ngram_tracing_and_supervised_ML.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ngram tracing \n",
    "## as introduced by Grieve et al. 2017\n",
    " make sure that all excel tables are in the same folder as this jupyter notebook if you intend to execute cells."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# imports packages\n",
    "import statistics as st\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import random\n",
    "import matplotlib.pyplot as plt\n",
    "#import frog\n",
    "from sklearn import svm\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.model_selection import train_test_split, KFold, cross_val_score, LeaveOneOut, ShuffleSplit\n",
    "from sklearn.metrics import accuracy_score, classification_report, confusion_matrix\n",
    "STOPWORDS = '.'\n",
    "df = pd.read_excel('df_with_lemmata.xlsx')\n",
    "df.set_index(\"ID\", inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Label</th>\n",
       "      <th>Taaloefening 1</th>\n",
       "      <th>Taaloefening 2</th>\n",
       "      <th>Text</th>\n",
       "      <th>Lemmas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>P642</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "      <td>terwijl de kok zijn deeg aan_het voorbereiden ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C021</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "      <td>je nemen de nodig hoeveelheid water , deze gie...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C022</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "      <td>de kok zijn een pizza aan_het maken in zijn re...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C041</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "      <td>op de prin zien je een pizza maker die een dee...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C042</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "      <td>je nemen een hoog bokaal , een filter . de fil...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Label                                     Taaloefening 1  \\\n",
       "ID                                                               \n",
       "P642      2                                                NaN   \n",
       "C021      0  Je neemt de nodige hoeveelheid water, deze gie...   \n",
       "C022      0                                                NaN   \n",
       "C041      0                                                NaN   \n",
       "C042      0  Je neemt een hoog bokaal, een filter. De filte...   \n",
       "\n",
       "                                         Taaloefening 2  \\\n",
       "ID                                                        \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...   \n",
       "C021                                                NaN   \n",
       "C022  De kok is een pizza aan het maken in zijn rest...   \n",
       "C041  Op de printjes zie je een pizza maker die een ...   \n",
       "C042                                                NaN   \n",
       "\n",
       "                                                   Text  \\\n",
       "ID                                                        \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...   \n",
       "C021  Je neemt de nodige hoeveelheid water, deze gie...   \n",
       "C022  De kok is een pizza aan het maken in zijn rest...   \n",
       "C041  Op de printjes zie je een pizza maker die een ...   \n",
       "C042  Je neemt een hoog bokaal, een filter. De filte...   \n",
       "\n",
       "                                                 Lemmas  \n",
       "ID                                                       \n",
       "P642  terwijl de kok zijn deeg aan_het voorbereiden ...  \n",
       "C021  je nemen de nodig hoeveelheid water , deze gie...  \n",
       "C022  de kok zijn een pizza aan_het maken in zijn re...  \n",
       "C041  op de prin zien je een pizza maker die een dee...  \n",
       "C042  je nemen een hoog bokaal , een filter . de fil...  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "textlength_c2 = []\n",
    "textlength_c1 = []\n",
    "c0_1 = []\n",
    "c0_2 = []\n",
    "\n",
    "\n",
    "pa_1 = []\n",
    "pa_2 = []\n",
    "textlength_pa1 = []\n",
    "textlength_pa2 = []\n",
    "pr_1 = []\n",
    "pr_2 = []\n",
    "textlength_pr1 = []\n",
    "textlength_pr2 = []\n",
    "\n",
    "\n",
    "for index, row in df.iterrows():\n",
    "    if df.loc[index, \"Label\"] == 0:\n",
    "        if isinstance(df.loc[index, \"Taaloefening 1\"], float) is not True:\n",
    "            c0_1.append(df.loc[index, \"Taaloefening 1\"])\n",
    "            textlength_c1.append(len(df.loc[index, \"Taaloefening 1\"]))\n",
    "        if isinstance(df.loc[index, \"Taaloefening 2\"], float) is not True:\n",
    "            c0_2.append(df.loc[index, \"Taaloefening 2\"])\n",
    "            textlength_c2.append(len(df.loc[index, \"Taaloefening 2\"]))\n",
    "            \n",
    "    if df.loc[index, \"Label\"] == 1:\n",
    "        if isinstance(df.loc[index, \"Taaloefening 1\"], float) is not True:\n",
    "            pa_1.append(df.loc[index, \"Taaloefening 1\"])\n",
    "            textlength_pa1.append(len(df.loc[index, \"Taaloefening 1\"]))\n",
    "\n",
    "        if isinstance(df.loc[index, \"Taaloefening 2\"], float) is not True:\n",
    "            pa_2.append(df.loc[index, \"Taaloefening 2\"])\n",
    "            textlength_pa2.append(len(df.loc[index, \"Taaloefening 2\"]))\n",
    "\n",
    "    if df.loc[index, \"Label\"] == 2:\n",
    "        if isinstance(df.loc[index, \"Taaloefening 1\"], float) is not True:\n",
    "            pr_1.append(df.loc[index, \"Taaloefening 1\"])\n",
    "            textlength_pr1.append(len(df.loc[index, \"Taaloefening 1\"]))\n",
    "\n",
    "        if isinstance(df.loc[index, \"Taaloefening 2\"], float) is not True:\n",
    "            pr_2.append(df.loc[index, \"Taaloefening 2\"])\n",
    "            textlength_pr2.append(len(df.loc[index, \"Taaloefening 2\"]))\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'frog' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-4-e9c780f00465>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     29\u001b[0m         \u001b[0mtext\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Taaloefening 2\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     30\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"Text\"\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtext\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 31\u001b[0;31m \u001b[0mfrog\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfrog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mFrog\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfrog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mFrogOptions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparser\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     32\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mlemmatize\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     33\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0;34m' '\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mword\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'lemma'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mword\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfrog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprocess\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mword\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'frog' is not defined"
     ]
    }
   ],
   "source": [
    "df = pd.read_excel('schizophrenia_data_cleaned.xlsx')\n",
    "# P662 exists twice, for Taaloefening 1 and 2 once\n",
    "df.loc[df.index == 76, \"Taaloefening 2\"] = df.loc[166][\"Taaloefening 2\"]\n",
    "# dropping irrelevant columns, LDP = \"Livia de Picker\", PANS = \"Positive and Negative Syndromes. Can be ignored due to absence of exact test scores\"\n",
    "columns = ['LDP', 'Datum', 'PANS', 'Unnamed: 5', 'Unnamed: 11', 'Unnamed: 13','1_woord_streep', '1_zin_streep', '1_alinea_streep', '1_Typografisch', '1_Comment', \"2_woord_streep\", \"2_zin_streep\",\"2_alinea_streep\", \"2_Typografisch\", \"2_Comment\",\"afterthought_word\",\"Spelcorrecties\",\"ALLCAPS\",\"UPPERCASE\",\"lowercase\",\"afterthought_phrase\"]\n",
    "df.drop(columns, inplace=True, axis=1)\n",
    "# add \"labels\" to dataframe; 0: control, 1: acute, 2: remission, 3: non-valid participant\n",
    "labels = []\n",
    "for index,row in df.iterrows():\n",
    "    if row[0].startswith('C'):\n",
    "        labels.append(0)\n",
    "    elif row[0].startswith('P') and row[0].endswith(\"1\"):\n",
    "        labels.append(1)\n",
    "    elif row[0].startswith('P') and row[0].endswith(\"2\"):\n",
    "        labels.append(2)\n",
    "    else:\n",
    "        df.drop(index, inplace=True, axis=0)\n",
    "df.insert(0, \"Label\", labels, True)\n",
    "df.set_index(\"ID\", inplace = True)\n",
    "# removes other duplicates of the data (e.g. P642)\n",
    "df = df.loc[~df.index.duplicated(keep='first')]\n",
    "# fixes a cell missing a 'nan' (- or data, but it's empty)\n",
    "df.loc[\"C092\", \"Taaloefening 2\"] = float('nan')\n",
    "text = []\n",
    "for index, row in df.iterrows():\n",
    "    if isinstance(df.loc[index, \"Taaloefening 1\"], float) is not True:\n",
    "        text.append(df.loc[index, \"Taaloefening 1\"])\n",
    "    elif isinstance(df.loc[index, \"Taaloefening 2\"], float) is not True:\n",
    "        text.append(df.loc[index, \"Taaloefening 2\"])\n",
    "df[\"Text\"] = text\n",
    "frog = frog.Frog(frog.FrogOptions(parser=False))\n",
    "def lemmatize(text):\n",
    "    return ' '.join([word['lemma'] for word in frog.process(text) if word])\n",
    "lemmas = []  \n",
    "for index, row in df.iterrows():\n",
    "    lemmas.append(lemmatize(df.loc[index, \"Text\"]))\n",
    "    \n",
    "df[\"Lemmas\"] = lemmas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## N-Gram-Tracing\n",
    "Defines all necessary functions for N-Gram-Tracing"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "# generate ngrams INCLUDING crossing sentence boundaries\n",
    "def generate_ngram(source, chars_or_words, ngram_order):\n",
    "    \"\"\"\n",
    "    Input: Source: string of words, chars_or_words: ngrams with \"characters\" or \"words\", ngram_order: integer (unigram, bigram, trigram,...)\n",
    "    Output: List with tuples of ngrams (strings)\n",
    "    \"\"\"\n",
    "    list_of_tuples = []\n",
    "    source = source.lower()\n",
    "    \n",
    "    if chars_or_words == \"words\":\n",
    "        source = source.split()\n",
    "        \n",
    "    for i in range(0, len(source) - ngram_order + 1): \n",
    "        ngram_order_index = i + ngram_order          \n",
    "        generated_ngram = tuple(source[i : ngram_order_index])\n",
    "        list_of_tuples.append(generated_ngram)   \n",
    "    return list_of_tuples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate ngrams WITHOUT crossing sentence boundaries\n",
    "def generate_ngram(source, chars_or_words, ngram_order):\n",
    "    \"\"\"\n",
    "    Input: Source: string of words, chars_or_words: ngrams with \"characters\" or \"words\", ngram_order: integer (unigram, bigram, trigram,...)\n",
    "    Output: List with tuples of ngrams (strings)\n",
    "    \"\"\"\n",
    "    list_of_tuples = []\n",
    "    output = []\n",
    "    source = source.lower()\n",
    "    \n",
    "    if chars_or_words == \"words\":\n",
    "        for word in source.split():\n",
    "            if '.' in word:\n",
    "                output.append(word[:-1])\n",
    "                output.append(\".\")\n",
    "            elif '.' not in word:\n",
    "                output.append(word)\n",
    "        source = output\n",
    "            \n",
    "    for i in range(0, len(source) - ngram_order + 1):\n",
    "        \n",
    "        ngram_order_index = i + ngram_order\n",
    "        generated_ngram = tuple(source[i : ngram_order_index])\n",
    "        if chars_or_words == \"chars\":\n",
    "            if not any(w in STOPWORDS for w in generated_ngram[:-1]):\n",
    "                list_of_tuples.append(generated_ngram)\n",
    "                \n",
    "        if chars_or_words == \"words\":\n",
    "            if not any(ngrams in STOPWORDS for ngrams in generated_ngram[:-1]):\n",
    "                list_of_tuples.append(generated_ngram)\n",
    "                \n",
    "    return list_of_tuples\n",
    "\n",
    "def overlapp_coefficient(set_of_ngrams1,set_of_ngrams2):\n",
    "    \"\"\"\n",
    "    Input: Two sets of ngrams that are tested against each other\n",
    "    Output: Overlapp Co-Efficient as described by Grieve et al. 2017:\n",
    "    > Measures the percentage of n-gram types found in the questioned document that also occur at least once in each possible author \n",
    "    (=> participant group) writing sample. <\n",
    "    \"\"\"\n",
    "    \n",
    "    return len(set(set_of_ngrams1).intersection(set(set_of_ngrams2))) / len(set(set_of_ngrams1))\n",
    "\n",
    "def create_dataframe(task, chars_or_words, ngram_order, amount_of_groups, participant_grouping):\n",
    "    \"\"\"\n",
    "    Creates a dataframe with all ngrams and random ngram selection based on whether the input was \"word\" or \"chars\", the ngram order\n",
    "    the amount of groups \"two_groups\" or \"three_groups\" and the constellation of participant_grouping: None, \"controls_vs_patients\" or \"controlsremission_vs_acute\" \n",
    "    \"\"\" \n",
    "    if task == 1:\n",
    "        return ngram_tracing_main(\"Taaloefening 1\", chars_or_words, ngram_order, amount_of_groups, participant_grouping)\n",
    "    if task == 2:\n",
    "        return ngram_tracing_main(\"Taaloefening 2\", chars_or_words, ngram_order, amount_of_groups, participant_grouping)\n",
    "\n",
    "    if task == \"both\":\n",
    "        df_ngrams_1 = ngram_tracing_main(\"Taaloefening 1\", chars_or_words, ngram_order, amount_of_groups, participant_grouping)\n",
    "        df_ngrams_2 = ngram_tracing_main(\"Taaloefening 2\", chars_or_words, ngram_order, amount_of_groups, participant_grouping)\n",
    "        return pd.concat([df_ngrams_1, df_ngrams_2], ignore_index=True)\n",
    "    \n",
    "def ngram_tracing_main(task, chars_or_words, ngram_order, amount_of_groups, participant_grouping):\n",
    "    \"\"\"\n",
    "    Input: takes the file to be tested (test_file), the task (Taaloefening 1 or 2), \n",
    "        whether ngrams should be made of characters or words (chars_or_words) and the ngram order\n",
    "    Output: creates a bag of ngrams for the testfile and each participant group, calculates the overlap coefficient for all 3 test cases, \n",
    "        chooses the highest and adds the results to a list of lists \"data\" which will be appended to the ngram dataframe later.\n",
    "    \"\"\"\n",
    "    data = []\n",
    "    for participant in df.index:\n",
    "        \n",
    "        ngrams0 = []\n",
    "        ngrams1 = []\n",
    "        ngrams2 = []\n",
    "        lemmasngrams0 = []\n",
    "        lemmasngrams1 = []\n",
    "        lemmasngrams2 = []\n",
    "        merged_label = 0\n",
    "        \n",
    "        if isinstance(df.loc[participant, task], float) is not True:\n",
    "            participant_ngrams = generate_ngram(df.loc[participant, task], chars_or_words, ngram_order)\n",
    "            if participant_grouping == \"controls_vs_patients\":\n",
    "                if df.loc[participant, \"Label\"] == 2:\n",
    "                    merged_label = 1\n",
    "                else:\n",
    "                    merged_label = df.loc[participant, \"Label\"]\n",
    "            elif participant_grouping == \"controlsremission_vs_acute\":\n",
    "                if df.loc[participant, \"Label\"] == 2:\n",
    "                    merged_label = 0\n",
    "                else:\n",
    "                    merged_label = df.loc[participant, \"Label\"]\n",
    "        \n",
    "            for participant2 in df.index:\n",
    "                if participant2 != participant:\n",
    "                \n",
    "                    if amount_of_groups == \"three_groups\":\n",
    "                        if df.loc[participant2, \"Label\"] == 0:\n",
    "                            if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                ngrams0 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                lemmasngrams0 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                        elif df.loc[participant2, \"Label\"] == 1:\n",
    "                            if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                ngrams1 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                lemmasngrams1 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                        elif df.loc[participant2, \"Label\"] == 2:\n",
    "                            if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                ngrams2 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                lemmasngrams2 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                    elif amount_of_groups == \"two_groups\":\n",
    "                        if participant_grouping == \"controls_vs_patients\":\n",
    "                            if df.loc[participant2, \"Label\"] == 0:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams0 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams0 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                            elif df.loc[participant2, \"Label\"] == 1:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams1 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams1 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                            elif df.loc[participant2, \"Label\"] == 2:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams1 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams1 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                        elif participant_grouping == \"controlsremission_vs_acute\":\n",
    "                            if df.loc[participant2, \"Label\"] == 0:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams0 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams0 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "                                    \n",
    "                            elif df.loc[participant2, \"Label\"] == 2:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams0 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams0 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "\n",
    "                            elif df.loc[participant2, \"Label\"] == 1:\n",
    "                                if isinstance(df.loc[participant2, task], float) is not True:\n",
    "                                    ngrams1 += generate_ngram(df.loc[participant2, task], chars_or_words, ngram_order)\n",
    "                                    lemmasngrams1 += generate_ngram(df.loc[participant2, \"Lemmas\"], chars_or_words, ngram_order)\n",
    "\n",
    "            \n",
    "            if amount_of_groups == \"three_groups\":\n",
    "                smallest_sample = min(len(ngrams0), len(ngrams1), len(ngrams2))\n",
    "                smallest_lemmasample = min(len(lemmasngrams0), len(lemmasngrams1), len(lemmasngrams2))\n",
    "            elif amount_of_groups == \"two_groups\":\n",
    "                smallest_sample = min(len(ngrams0), len(ngrams1))\n",
    "                smallest_lemmasample = min(len(lemmasngrams0), len(lemmasngrams1))\n",
    "                \n",
    "            random0 = random.sample(ngrams0, smallest_sample)\n",
    "            random1 = random.sample(ngrams1, smallest_sample)\n",
    "            randomlemma0 = random.sample(lemmasngrams0, smallest_lemmasample)\n",
    "            randomlemma1 = random.sample(lemmasngrams1, smallest_lemmasample)\n",
    "\n",
    "            ngrams0_ratio = overlapp_coefficient(participant_ngrams,random0)\n",
    "            ngrams1_ratio = overlapp_coefficient(participant_ngrams,random1)\n",
    "            lemmas0_ratio = overlapp_coefficient(participant_ngrams,randomlemma0)\n",
    "            lemmas1_ratio = overlapp_coefficient(participant_ngrams,randomlemma1)\n",
    "        \n",
    "            if amount_of_groups == \"two_groups\":\n",
    "                ratio_list = [ngrams0_ratio, ngrams1_ratio]\n",
    "                lemmaratio_list = [lemmas0_ratio, lemmas1_ratio]\n",
    "                overlap_difference = abs(ngrams0_ratio - ngrams1_ratio)\n",
    "                lemma_overlap_difference = abs(lemmas0_ratio - lemmas1_ratio)\n",
    "\n",
    "            elif amount_of_groups == \"three_groups\":\n",
    "                random2 = random.sample(ngrams2, smallest_sample)\n",
    "                randomlemma2 = random.sample(lemmasngrams2, smallest_lemmasample)\n",
    "                ngrams2_ratio = overlapp_coefficient(participant_ngrams,random2)\n",
    "                lemmas2_ratio = overlapp_coefficient(participant_ngrams,randomlemma2)\n",
    "                ratio_list = [ngrams0_ratio, ngrams1_ratio, ngrams2_ratio]\n",
    "                lemmaratio_list = [lemmas0_ratio, lemmas1_ratio, lemmas2_ratio]\n",
    "                difference_list = ratio_list.copy()\n",
    "                difference_list_lemmas = lemmaratio_list.copy()\n",
    "                minimum = min(difference_list)\n",
    "                minimumlemmas = min(difference_list_lemmas)\n",
    "                difference_list.remove(minimum)\n",
    "                difference_list_lemmas.remove(minimumlemmas)\n",
    "                overlap_difference = abs(difference_list[0] - difference_list[1])\n",
    "                overlap_difference_lemmas = abs(difference_list_lemmas[0] - difference_list_lemmas[1])\n",
    "            \n",
    "            if max(ratio_list) == 0.0:\n",
    "                predicted_label = float('nan')\n",
    "                predicted_label_lemmas = lemmaratio_list.index(max(lemmaratio_list))\n",
    "            else:\n",
    "                predicted_label = ratio_list.index(max(ratio_list))\n",
    "                predicted_label_lemmas = lemmaratio_list.index(max(lemmaratio_list))\n",
    "        \n",
    "            if amount_of_groups == \"three_groups\":\n",
    "                data.append([participant, df.loc[participant, \"Label\"], predicted_label,predicted_label_lemmas,  overlap_difference, ngrams0_ratio, ngrams1_ratio, ngrams2_ratio, participant_ngrams, random0, random1, random2, ngrams0, ngrams1, ngrams2])\n",
    "                df_ngrams = pd.DataFrame(data, columns = ['Participant', 'Label', 'Predicted Label', \"Predicted Label Lemmas\", \"Overlap Difference\",  'Overlap with Label 0', 'Overlap with Label 1', \"Overlap with Label 2\", \"Participant Ngrams\", \"Random 0\", \"Random 1\", \"Random 2\", \"Ngrams 0\", \"Ngrams 1\", \"Ngrams 2\"])\n",
    "\n",
    "            elif amount_of_groups == \"two_groups\":\n",
    "                data.append([participant, merged_label, predicted_label, predicted_label_lemmas, overlap_difference, ngrams0_ratio, ngrams1_ratio, participant_ngrams, random0, random1, ngrams0, ngrams1]) \n",
    "                df_ngrams = pd.DataFrame(data, columns = ['Participant', 'Label', 'Predicted Label', \"Predicted Label Lemmas\", \"Overlap Difference\",  'Overlap with Label 0', 'Overlap with Label 1', \"Participant Ngrams\", \"Random 0\", \"Random 1\", \"Ngrams 0\", \"Ngrams 1\"])\n",
    "    return df_ngrams\n",
    "\n",
    "def accuracy(dataframe):\n",
    "    \"\"\"\n",
    "    Prints the Accuracy, the amount of correctly predicted participants and the amount of all participants\n",
    "    \"\"\"\n",
    "    correctly_predicted = 0\n",
    "    for index,row in dataframe.iterrows():\n",
    "        if row[1] == row[2]:\n",
    "            correctly_predicted += 1\n",
    "    accuracy = correctly_predicted / len(dataframe)\n",
    "    ac_counter.append(accuracy)\n",
    "    correctly_predicted_lemmas = 0\n",
    "    for index,row in dataframe.iterrows():\n",
    "        if row[1] == row[3]:\n",
    "            correctly_predicted_lemmas += 1\n",
    "    accuracy_lemmas = correctly_predicted_lemmas / len(dataframe)\n",
    "    le_counter.append(accuracy_lemmas)\n",
    "    #print(\"Correctly predicted:\", correctly_predicted, \"/\", len(dataframe))\n",
    "    #print(\"Accuracy:\", accuracy, \"Accuracy with Lemmas\", accuracy_lemmas, \"\\n\",)\n",
    "    return(ac_counter, le_counter)\n",
    "\n",
    "def supportvectormachine(X, y, X_train, y_train):\n",
    "    SVM = SVC(kernel='linear')\n",
    "    SVM.fit(X_train, y_train)\n",
    "    y_pred = SVM.predict(X_test)\n",
    "    scores = cross_val_score(SVM, X, y, cv=5)\n",
    "    print(\"---------- Support Vector Machine ----------\",\"\\n\")\n",
    "    return SVM, y_pred, scores\n",
    "\n",
    "def naivebayes(X, y, X_train, y_train):\n",
    "    NB = MultinomialNB()\n",
    "    NB.fit(X_train, y_train)\n",
    "    y_pred = NB.predict(X_test)\n",
    "    scores = cross_val_score(NB, X, y, cv=5)\n",
    "    print(\"---------- Naive Bayes ----------\")\n",
    "    return NB, y_pred, scores\n",
    "\n",
    "def classifierresults(X, y, y_test, y_pred, scores, model):\n",
    "    print(\"Confusion Matrix:\",\"\\n\",confusion_matrix(y_test,y_pred))\n",
    "    print(classification_report(y_test,y_pred))\n",
    "    print(\"Accuracy: %0.2f (+/- %0.2f)\" % (scores.mean(), scores.std() * 2))\n",
    "    seed = 7\n",
    "    test_size = 0.33\n",
    "    kfold = KFold(n_splits=25, random_state=seed)\n",
    "    results = cross_val_score(model, X, y, cv=kfold)\n",
    "    print(\"K-Fold Cross Validation: Accuracy: %.3f%% (+/- %.3f%%)\" % (results.mean()*100.0, results.std()*100.0))\n",
    "    loocv = LeaveOneOut()\n",
    "    results = cross_val_score(model, X, y, cv=loocv)\n",
    "    print(\"Leave-One-Out Cross Validation: Accuracy: %.3f%% (+/- %.3f%%)\" % (results.mean()*100.0, results.std()*100.0))\n",
    "    kfold = ShuffleSplit(n_splits=10, test_size=test_size, random_state=seed)\n",
    "    results = cross_val_score(model, X, y, cv=kfold)\n",
    "    print(\"Repeated Random Test-Train Splits: Accuracy: %.3f%% (+/- %.3f%%)\" % (results.mean()*100.0, results.std()*100.0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The following cells create a dataframe with the ngram tracing data with one row per participant. It calculates the accuracy, confusion matrix and prints some cases of \"clear decision\"\n",
    "add a new cell with \"df_ngrams\" or \"df_ngrams.head()\" to have a closer look at the dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.4544303797468355\n",
      "[0.379746835443038, 0.4936708860759494, 0.5063291139240507, 0.4810126582278481, 0.3670886075949367, 0.43037974683544306, 0.4430379746835443, 0.46835443037974683, 0.4936708860759494, 0.4810126582278481]\n",
      "0.4341772151898734\n",
      "[0.5063291139240507, 0.45569620253164556, 0.3670886075949367, 0.3670886075949367, 0.43037974683544306, 0.3924050632911392, 0.4936708860759494, 0.4177215189873418, 0.4177215189873418, 0.4936708860759494]\n"
     ]
    }
   ],
   "source": [
    "# Taaloefening 1 with word ngrams, ngram order 1 and the three groups control vs acute vs remission\n",
    "# parameters for \"two_groups\": controls_vs_patients controlsremission_vs_acute\n",
    "\n",
    "ac_counter = []\n",
    "le_counter = []\n",
    "for _ in range(10):\n",
    "    df_ngrams = create_dataframe(both,\"chars\", 2, \"three_groups\", participant_grouping = None)\n",
    "    y_test = df_ngrams[\"Label\"]; y_pred = df_ngrams[\"Predicted Label\"]\n",
    "    accuracy(df_ngrams)\n",
    "    \n",
    "    # have the classification report and confusion matrix additionally printed to each iteration\n",
    "    #print(classification_report(y_test, y_pred))\n",
    "    #print(confusion_matrix(y_test, y_pred))\n",
    "    \n",
    "print(\"Average Accuracy over all iterations for raw words <ac>\":, sum(ac_counter) / len(ac_counter))\n",
    "print(ac_counter)\n",
    "print(\"Average Accuracy over all iterations for lemmata <le>\"sum(le_counter) / len(le_counter))\n",
    "print(le_counter)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## The following 3 cells manually store the information from ALL ngram accuracies (DON'T DELETE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "le1words_1 = [0.40476190476190477, 0.30952380952380953, 0.36904761904761907, 0.2857142857142857, 0.3333333333333333, 0.3333333333333333, 0.36904761904761907, 0.32142857142857145, 0.32142857142857145, 0.35714285714285715]\n",
    "le2words_1 = [0.21428571428571427, 0.2976190476190476, 0.21428571428571427, 0.27380952380952384, 0.25, 0.2619047619047619, 0.2976190476190476, 0.2857142857142857, 0.32142857142857145, 0.27380952380952384]\n",
    "le3words_1 = [0.4642857142857143, 0.5119047619047619, 0.4642857142857143, 0.40476190476190477, 0.4880952380952381, 0.5595238095238095, 0.4166666666666667, 0.5238095238095238, 0.5357142857142857, 0.5357142857142857]\n",
    "le3chars_1 = [0.35714285714285715, 0.4166666666666667, 0.36904761904761907, 0.32142857142857145, 0.39285714285714285, 0.4166666666666667, 0.38095238095238093, 0.47619047619047616, 0.47619047619047616, 0.4166666666666667]\n",
    "le4chars_1 = [0.35714285714285715, 0.44047619047619047, 0.44047619047619047, 0.34523809523809523, 0.2857142857142857, 0.35714285714285715, 0.32142857142857145, 0.2976190476190476, 0.38095238095238093, 0.35714285714285715]\n",
    "le5chars_1 = [0.2857142857142857, 0.30952380952380953, 0.32142857142857145, 0.42857142857142855, 0.2619047619047619, 0.4166666666666667, 0.35714285714285715, 0.30952380952380953, 0.36904761904761907, 0.3333333333333333]\n",
    "le6chars_1 = [0.23809523809523808, 0.2619047619047619, 0.2976190476190476, 0.2261904761904762, 0.20238095238095238, 0.20238095238095238, 0.2857142857142857, 0.3333333333333333, 0.27380952380952384, 0.25]\n",
    "le7chars_1 = [0.2619047619047619, 0.27380952380952384, 0.2619047619047619, 0.21428571428571427, 0.21428571428571427, 0.2619047619047619, 0.2261904761904762, 0.2619047619047619, 0.2261904761904762, 0.2619047619047619]\n",
    "le8chars_1 = [0.21428571428571427, 0.23809523809523808, 0.21428571428571427, 0.2857142857142857, 0.2976190476190476, 0.17857142857142858, 0.17857142857142858, 0.2261904761904762, 0.21428571428571427, 0.2619047619047619]\n",
    "le9chars_1 = [0.21428571428571427, 0.23809523809523808, 0.2619047619047619, 0.2857142857142857, 0.25, 0.2619047619047619, 0.23809523809523808, 0.20238095238095238, 0.27380952380952384, 0.2857142857142857]\n",
    "le10chars_1 = [0.27380952380952384, 0.2619047619047619, 0.2261904761904762, 0.25, 0.25, 0.2619047619047619, 0.20238095238095238, 0.23809523809523808, 0.27380952380952384, 0.3333333333333333]\n",
    "le11chars_1 = [0.39285714285714285, 0.2976190476190476, 0.30952380952380953, 0.23809523809523808, 0.2976190476190476, 0.3333333333333333, 0.2857142857142857, 0.30952380952380953, 0.35714285714285715, 0.27380952380952384]\n",
    "le12chars_1 = [0.23809523809523808, 0.2261904761904762, 0.2976190476190476, 0.2857142857142857, 0.34523809523809523, 0.30952380952380953, 0.3333333333333333, 0.3333333333333333, 0.32142857142857145, 0.35714285714285715]\n",
    "le13chars_1 = [0.32142857142857145, 0.30952380952380953, 0.30952380952380953, 0.2857142857142857, 0.32142857142857145, 0.27380952380952384, 0.25, 0.3333333333333333, 0.39285714285714285, 0.32142857142857145]\n",
    "le14chars_1 = [0.23809523809523808, 0.3333333333333333, 0.32142857142857145, 0.30952380952380953, 0.32142857142857145, 0.32142857142857145, 0.23809523809523808, 0.39285714285714285, 0.32142857142857145, 0.3333333333333333]\n",
    "le15chars_1 = [0.30952380952380953, 0.2976190476190476, 0.30952380952380953, 0.36904761904761907, 0.3333333333333333, 0.35714285714285715, 0.34523809523809523, 0.25, 0.36904761904761907, 0.34523809523809523]\n",
    "le16chars_1 = [0.32142857142857145, 0.27380952380952384, 0.4166666666666667, 0.2857142857142857, 0.34523809523809523, 0.38095238095238093, 0.3333333333333333, 0.34523809523809523, 0.32142857142857145, 0.3333333333333333]\n",
    "ac1words_1 = [0.34523809523809523, 0.32142857142857145, 0.3333333333333333, 0.2976190476190476, 0.40476190476190477, 0.35714285714285715, 0.34523809523809523, 0.4166666666666667, 0.35714285714285715, 0.35714285714285715]\n",
    "ac2words_1 = [0.2976190476190476, 0.2976190476190476, 0.35714285714285715, 0.35714285714285715, 0.30952380952380953, 0.2619047619047619, 0.2619047619047619, 0.30952380952380953, 0.2976190476190476, 0.2976190476190476]\n",
    "ac3words_1 = [0.38095238095238093, 0.34523809523809523, 0.3333333333333333, 0.40476190476190477, 0.27380952380952384, 0.40476190476190477, 0.4166666666666667, 0.42857142857142855, 0.3333333333333333, 0.38095238095238093]\n",
    "ac3chars_1 = [0.4642857142857143, 0.47619047619047616, 0.40476190476190477, 0.44047619047619047, 0.4642857142857143, 0.4166666666666667, 0.42857142857142855, 0.34523809523809523, 0.42857142857142855, 0.42857142857142855]\n",
    "ac4chars_1 = [0.3333333333333333, 0.38095238095238093, 0.40476190476190477, 0.34523809523809523, 0.32142857142857145, 0.47619047619047616, 0.4166666666666667, 0.34523809523809523, 0.38095238095238093, 0.40476190476190477]\n",
    "ac5chars_1 = [0.27380952380952384, 0.2619047619047619, 0.2976190476190476, 0.30952380952380953, 0.2976190476190476, 0.39285714285714285, 0.34523809523809523, 0.27380952380952384, 0.32142857142857145, 0.30952380952380953]\n",
    "ac6chars_1 = [0.2857142857142857, 0.2976190476190476, 0.30952380952380953, 0.32142857142857145, 0.2976190476190476, 0.25, 0.34523809523809523, 0.32142857142857145, 0.34523809523809523, 0.25]\n",
    "ac7chars_1 = [0.2857142857142857, 0.2619047619047619, 0.2619047619047619, 0.2857142857142857, 0.30952380952380953, 0.25, 0.23809523809523808, 0.27380952380952384, 0.25, 0.23809523809523808]\n",
    "ac8chars_1 = [0.2261904761904762, 0.2619047619047619, 0.2976190476190476, 0.2976190476190476, 0.20238095238095238, 0.2976190476190476, 0.2261904761904762, 0.2619047619047619, 0.2976190476190476, 0.2976190476190476]\n",
    "ac9chars_1 = [0.30952380952380953, 0.27380952380952384, 0.2857142857142857, 0.2857142857142857, 0.32142857142857145, 0.2261904761904762, 0.2857142857142857, 0.30952380952380953, 0.34523809523809523, 0.27380952380952384]\n",
    "ac10chars_1 = [0.32142857142857145, 0.27380952380952384, 0.2976190476190476, 0.38095238095238093, 0.3333333333333333, 0.34523809523809523, 0.30952380952380953, 0.34523809523809523, 0.2976190476190476, 0.32142857142857145]\n",
    "ac11chars_1 = [0.32142857142857145, 0.30952380952380953, 0.3333333333333333, 0.34523809523809523, 0.32142857142857145, 0.36904761904761907, 0.30952380952380953, 0.32142857142857145, 0.34523809523809523, 0.30952380952380953]\n",
    "ac12chars_1 = [0.34523809523809523, 0.38095238095238093, 0.44047619047619047, 0.34523809523809523, 0.38095238095238093, 0.34523809523809523, 0.39285714285714285, 0.35714285714285715, 0.39285714285714285, 0.34523809523809523]\n",
    "ac13chars_1 = [0.4166666666666667, 0.3333333333333333, 0.2857142857142857, 0.35714285714285715, 0.35714285714285715, 0.4166666666666667, 0.42857142857142855, 0.35714285714285715, 0.35714285714285715, 0.38095238095238093]\n",
    "ac14chars_1 = [0.38095238095238093, 0.35714285714285715, 0.27380952380952384, 0.36904761904761907, 0.23809523809523808, 0.42857142857142855, 0.34523809523809523, 0.39285714285714285, 0.30952380952380953, 0.35714285714285715]\n",
    "ac15chars_1 = [0.30952380952380953, 0.2976190476190476, 0.32142857142857145, 0.34523809523809523, 0.34523809523809523, 0.39285714285714285, 0.40476190476190477, 0.4166666666666667, 0.42857142857142855, 0.34523809523809523]\n",
    "ac16chars_1 = [0.40476190476190477, 0.2976190476190476, 0.36904761904761907, 0.32142857142857145, 0.2857142857142857, 0.32142857142857145, 0.25, 0.32142857142857145, 0.39285714285714285, 0.35714285714285715]\n",
    "# two groups\n",
    "con_acurem_le1words_1 = [0.44047619047619047, 0.5, 0.5714285714285714, 0.5476190476190477, 0.4642857142857143, 0.5, 0.5476190476190477, 0.5476190476190477, 0.5238095238095238, 0.5119047619047619]\n",
    "con_acurem_le2words_1 = [0.5119047619047619, 0.6547619047619048, 0.5595238095238095, 0.5833333333333334, 0.4880952380952381, 0.5714285714285714, 0.6190476190476191, 0.5833333333333334, 0.5714285714285714, 0.5119047619047619]\n",
    "con_acurem_le3words_1 = [0.5476190476190477, 0.5833333333333334, 0.4880952380952381, 0.5595238095238095, 0.5476190476190477, 0.5595238095238095, 0.5357142857142857, 0.5833333333333334, 0.5952380952380952, 0.5357142857142857]\n",
    "con_acurem_le3chars_1 = [0.5595238095238095, 0.5595238095238095, 0.5238095238095238, 0.6071428571428571, 0.6547619047619048, 0.6309523809523809, 0.5238095238095238, 0.5833333333333334, 0.5595238095238095, 0.5714285714285714]\n",
    "con_acurem_le4chars_1 = [0.6071428571428571, 0.5833333333333334, 0.6071428571428571, 0.5476190476190477, 0.5714285714285714, 0.5952380952380952, 0.5714285714285714, 0.5714285714285714, 0.5357142857142857, 0.5595238095238095]\n",
    "con_acurem_le5chars_1 = [0.6071428571428571, 0.5595238095238095, 0.5833333333333334, 0.6428571428571429, 0.5833333333333334, 0.5952380952380952, 0.6190476190476191, 0.5595238095238095, 0.6190476190476191, 0.5833333333333334]\n",
    "con_acurem_le6chars_1 = [0.5952380952380952, 0.5714285714285714, 0.6071428571428571, 0.6785714285714286, 0.6190476190476191, 0.6190476190476191, 0.6071428571428571, 0.5833333333333334, 0.6428571428571429, 0.6071428571428571]\n",
    "con_acurem_le7chars_1 = [0.6190476190476191, 0.6190476190476191, 0.6666666666666666, 0.6309523809523809, 0.6190476190476191, 0.6547619047619048, 0.6309523809523809, 0.5952380952380952, 0.5952380952380952, 0.6666666666666666]\n",
    "con_acurem_le8chars_1 = [0.6071428571428571, 0.6547619047619048, 0.6190476190476191, 0.5714285714285714, 0.6785714285714286, 0.6071428571428571, 0.5833333333333334, 0.6071428571428571, 0.6190476190476191, 0.5952380952380952]\n",
    "con_acurem_le9chars_1 = [0.6309523809523809, 0.6190476190476191, 0.5833333333333334, 0.5595238095238095, 0.5357142857142857, 0.5714285714285714, 0.6309523809523809, 0.5952380952380952, 0.5833333333333334, 0.6190476190476191]\n",
    "con_acurem_le10chars_1 = [0.5714285714285714, 0.5595238095238095, 0.5476190476190477, 0.5595238095238095, 0.5714285714285714, 0.5595238095238095, 0.6071428571428571, 0.5476190476190477, 0.5119047619047619, 0.5238095238095238]\n",
    "con_acurem_le11chars_1 = [0.5476190476190477, 0.5476190476190477, 0.5833333333333334, 0.5119047619047619, 0.5595238095238095, 0.5833333333333334, 0.5833333333333334, 0.5357142857142857, 0.5238095238095238, 0.5357142857142857]\n",
    "con_acurem_le12chars_1 = [0.5952380952380952, 0.5714285714285714, 0.5357142857142857, 0.5595238095238095, 0.5714285714285714, 0.6071428571428571, 0.6071428571428571, 0.5714285714285714, 0.5714285714285714, 0.5833333333333334]\n",
    "con_acurem_le13chars_1 = [0.6190476190476191, 0.5833333333333334, 0.5714285714285714, 0.6071428571428571, 0.6428571428571429, 0.6071428571428571, 0.6071428571428571, 0.6071428571428571, 0.6071428571428571, 0.5595238095238095]\n",
    "con_acurem_le14chars_1 = [0.6190476190476191, 0.5833333333333334, 0.5952380952380952, 0.6190476190476191, 0.5833333333333334, 0.6071428571428571, 0.5476190476190477, 0.5714285714285714, 0.5952380952380952, 0.5833333333333334]\n",
    "con_acurem_le15chars_1 = [0.5714285714285714, 0.5952380952380952, 0.5833333333333334, 0.6428571428571429, 0.6071428571428571, 0.6071428571428571, 0.5952380952380952, 0.5952380952380952, 0.5952380952380952, 0.6071428571428571]\n",
    "con_acurem_le16chars_1 = [0.6071428571428571, 0.6309523809523809, 0.5952380952380952, 0.6428571428571429, 0.6071428571428571, 0.5833333333333334, 0.5833333333333334, 0.6071428571428571, 0.5714285714285714, 0.6309523809523809]\n",
    "con_acurem_ac1words_1 = [0.5476190476190477, 0.5, 0.5238095238095238, 0.5357142857142857, 0.5476190476190477, 0.5357142857142857, 0.5476190476190477, 0.5833333333333334, 0.5714285714285714, 0.5595238095238095]\n",
    "con_acurem_ac2words_1 = [0.6190476190476191, 0.5714285714285714, 0.5714285714285714, 0.5595238095238095, 0.6547619047619048, 0.5714285714285714, 0.5714285714285714, 0.6309523809523809, 0.5476190476190477, 0.6071428571428571]\n",
    "con_acurem_ac3words_1 = [0.6190476190476191, 0.5952380952380952, 0.5952380952380952, 0.5952380952380952, 0.6190476190476191, 0.6071428571428571, 0.6071428571428571, 0.5833333333333334, 0.5833333333333334, 0.6309523809523809]\n",
    "con_acurem_ac3chars_1 = [0.6666666666666666, 0.5238095238095238, 0.5595238095238095, 0.5833333333333334, 0.5833333333333334, 0.5833333333333334, 0.5952380952380952, 0.5952380952380952, 0.5952380952380952, 0.5714285714285714]\n",
    "con_acurem_ac4chars_1 = [0.5952380952380952, 0.5833333333333334, 0.6547619047619048, 0.5833333333333334, 0.6428571428571429, 0.6190476190476191, 0.5476190476190477, 0.6190476190476191, 0.6666666666666666, 0.6071428571428571]\n",
    "con_acurem_ac5chars_1 = [0.6190476190476191, 0.5952380952380952, 0.6309523809523809, 0.6071428571428571, 0.6904761904761905, 0.5952380952380952, 0.6428571428571429, 0.7023809523809523, 0.6666666666666666, 0.6547619047619048]\n",
    "con_acurem_ac6chars_1 = [0.6904761904761905, 0.6666666666666666, 0.6190476190476191, 0.6666666666666666, 0.6547619047619048, 0.6547619047619048, 0.6547619047619048, 0.6428571428571429, 0.6309523809523809, 0.6547619047619048]\n",
    "con_acurem_ac7chars_1 = [0.6309523809523809, 0.6547619047619048, 0.6071428571428571, 0.6309523809523809, 0.6428571428571429, 0.6309523809523809, 0.6547619047619048, 0.6666666666666666, 0.6547619047619048, 0.6785714285714286]\n",
    "con_acurem_ac8chars_1 = [0.6785714285714286, 0.6547619047619048, 0.6785714285714286, 0.6428571428571429, 0.6904761904761905, 0.6666666666666666, 0.6309523809523809, 0.6904761904761905, 0.6666666666666666, 0.6428571428571429]\n",
    "con_acurem_ac9chars_1 = [0.6904761904761905, 0.6547619047619048, 0.6666666666666666, 0.6428571428571429, 0.6428571428571429, 0.6547619047619048, 0.6190476190476191, 0.6428571428571429, 0.6666666666666666, 0.6547619047619048]\n",
    "con_acurem_ac10chars_1 = [0.6190476190476191, 0.6428571428571429, 0.6785714285714286, 0.6547619047619048, 0.5952380952380952, 0.6309523809523809, 0.6309523809523809, 0.6309523809523809, 0.6666666666666666, 0.6666666666666666]\n",
    "con_acurem_ac11chars_1 = [0.6309523809523809, 0.6190476190476191, 0.5833333333333334, 0.6428571428571429, 0.6071428571428571, 0.6428571428571429, 0.6904761904761905, 0.6428571428571429, 0.6190476190476191, 0.6428571428571429]\n",
    "con_acurem_ac12chars_1 = [0.6309523809523809, 0.6190476190476191, 0.6547619047619048, 0.6071428571428571, 0.5952380952380952, 0.6547619047619048, 0.6666666666666666, 0.6190476190476191, 0.6190476190476191, 0.5595238095238095]\n",
    "con_acurem_ac13chars_1 = [0.6190476190476191, 0.6190476190476191, 0.6309523809523809, 0.6428571428571429, 0.6190476190476191, 0.6309523809523809, 0.5833333333333334, 0.6071428571428571, 0.6666666666666666, 0.6785714285714286]\n",
    "con_acurem_ac14chars_1 = [0.6547619047619048, 0.6785714285714286, 0.6785714285714286, 0.6071428571428571, 0.6547619047619048, 0.6666666666666666, 0.6547619047619048, 0.6547619047619048, 0.6547619047619048, 0.6309523809523809]\n",
    "con_acurem_ac15chars_1 = [0.6428571428571429, 0.6309523809523809, 0.6190476190476191, 0.6428571428571429, 0.6428571428571429, 0.6071428571428571, 0.5952380952380952, 0.6428571428571429, 0.5952380952380952, 0.6428571428571429]\n",
    "con_acurem_ac16chars_1 = [0.6666666666666666, 0.6309523809523809, 0.6428571428571429, 0.6309523809523809, 0.6309523809523809, 0.6428571428571429, 0.6190476190476191, 0.6666666666666666, 0.6428571428571429, 0.6666666666666666]\n",
    "\n",
    "conrem_acu_ac1words_1 = [0.47619047619047616, 0.5595238095238095, 0.5357142857142857, 0.5357142857142857, 0.5357142857142857, 0.4523809523809524, 0.47619047619047616, 0.5476190476190477, 0.4880952380952381, 0.5595238095238095]\n",
    "conrem_acu_ac2words_1 = [0.6071428571428571, 0.5476190476190477, 0.6309523809523809, 0.5833333333333334, 0.6309523809523809, 0.6071428571428571, 0.6547619047619048, 0.6428571428571429, 0.6190476190476191, 0.6190476190476191]\n",
    "conrem_acu_ac3words_1 = [0.6428571428571429, 0.5952380952380952, 0.5833333333333334, 0.5357142857142857, 0.5952380952380952, 0.5357142857142857, 0.5714285714285714, 0.5595238095238095, 0.6071428571428571, 0.5952380952380952]\n",
    "conrem_acu_ac3chars_1 = [0.6666666666666666, 0.6190476190476191, 0.6071428571428571, 0.6190476190476191, 0.5595238095238095, 0.6428571428571429, 0.6428571428571429, 0.6547619047619048, 0.6190476190476191, 0.5833333333333334]\n",
    "conrem_acu_ac4chars_1 = [0.6071428571428571, 0.6309523809523809, 0.6190476190476191, 0.5952380952380952, 0.6547619047619048, 0.6071428571428571, 0.5714285714285714, 0.5952380952380952, 0.6071428571428571, 0.6190476190476191]\n",
    "conrem_acu_ac5chars_1 = [0.6428571428571429, 0.5952380952380952, 0.5952380952380952, 0.6071428571428571, 0.6428571428571429, 0.6666666666666666, 0.5952380952380952, 0.5952380952380952, 0.5833333333333334, 0.6071428571428571]\n",
    "conrem_acu_ac6chars_1 = [0.6666666666666666, 0.6309523809523809, 0.6428571428571429, 0.6904761904761905, 0.7023809523809523, 0.7023809523809523, 0.6190476190476191, 0.6428571428571429, 0.6547619047619048, 0.6428571428571429]\n",
    "conrem_acu_ac7chars_1 = [0.7380952380952381, 0.7261904761904762, 0.6547619047619048, 0.6666666666666666, 0.6309523809523809, 0.6666666666666666, 0.6547619047619048, 0.6547619047619048, 0.6904761904761905, 0.6904761904761905]\n",
    "conrem_acu_ac8chars_1 = [0.6785714285714286, 0.6428571428571429, 0.6309523809523809, 0.6785714285714286, 0.7023809523809523, 0.7142857142857143, 0.6904761904761905, 0.6785714285714286, 0.6309523809523809, 0.6428571428571429]\n",
    "conrem_acu_ac9chars_1 = [0.6785714285714286, 0.6904761904761905, 0.6785714285714286, 0.6547619047619048, 0.6785714285714286, 0.6666666666666666, 0.7261904761904762, 0.6666666666666666, 0.6666666666666666, 0.6428571428571429]\n",
    "conrem_acu_ac10chars_1 = [0.6071428571428571, 0.6190476190476191, 0.6666666666666666, 0.6071428571428571, 0.6547619047619048, 0.6904761904761905, 0.6666666666666666, 0.6428571428571429, 0.6547619047619048, 0.6904761904761905]\n",
    "conrem_acu_ac11chars_1 = [0.6547619047619048, 0.6428571428571429, 0.6428571428571429, 0.6785714285714286, 0.6666666666666666, 0.6666666666666666, 0.6547619047619048, 0.6666666666666666, 0.6666666666666666, 0.6785714285714286]\n",
    "conrem_acu_ac12chars_1 = [0.6666666666666666, 0.6428571428571429, 0.6666666666666666, 0.6547619047619048, 0.6785714285714286, 0.6547619047619048, 0.6547619047619048, 0.6666666666666666, 0.6309523809523809, 0.6309523809523809]\n",
    "conrem_acu_ac13chars_1 = [0.6785714285714286, 0.6190476190476191, 0.6666666666666666, 0.6190476190476191, 0.6428571428571429, 0.6785714285714286, 0.6666666666666666, 0.6309523809523809, 0.6428571428571429, 0.7261904761904762]\n",
    "conrem_acu_ac14chars_1 = [0.6666666666666666, 0.6785714285714286, 0.6666666666666666, 0.6666666666666666, 0.6785714285714286, 0.6428571428571429, 0.6904761904761905, 0.6904761904761905, 0.6904761904761905, 0.7142857142857143]\n",
    "conrem_acu_ac15chars_1 = [0.7023809523809523, 0.6785714285714286, 0.6547619047619048, 0.7023809523809523, 0.6666666666666666, 0.6666666666666666, 0.6904761904761905, 0.6785714285714286, 0.6785714285714286, 0.6547619047619048]\n",
    "conrem_acu_ac16chars_1 = [0.6904761904761905, 0.6428571428571429, 0.6785714285714286, 0.6785714285714286, 0.6904761904761905, 0.6666666666666666, 0.6666666666666666, 0.6309523809523809, 0.6666666666666666, 0.6785714285714286]\n",
    "conrem_acu_le1words_1 = [0.5714285714285714, 0.5, 0.5119047619047619, 0.5119047619047619, 0.5595238095238095, 0.4880952380952381, 0.5357142857142857, 0.5238095238095238, 0.5595238095238095, 0.4523809523809524]\n",
    "conrem_acu_le2words_1 = [0.6071428571428571, 0.6309523809523809, 0.5476190476190477, 0.5833333333333334, 0.5476190476190477, 0.6071428571428571, 0.5952380952380952, 0.5357142857142857, 0.6666666666666666, 0.5833333333333334]\n",
    "conrem_acu_le3words_1 = [0.5357142857142857, 0.5, 0.5714285714285714, 0.5119047619047619, 0.5119047619047619, 0.5833333333333334, 0.5, 0.5238095238095238, 0.5119047619047619, 0.5714285714285714]\n",
    "conrem_acu_le3chars_1 = [0.5714285714285714, 0.6071428571428571, 0.5357142857142857, 0.5595238095238095, 0.5833333333333334, 0.5595238095238095, 0.5714285714285714, 0.5833333333333334, 0.5357142857142857, 0.5952380952380952]\n",
    "conrem_acu_le4chars_1 = [0.5833333333333334, 0.6428571428571429, 0.5595238095238095, 0.5357142857142857, 0.5, 0.5238095238095238, 0.5476190476190477, 0.5357142857142857, 0.5952380952380952, 0.5595238095238095]\n",
    "conrem_acu_le5chars_1 = [0.5833333333333334, 0.5595238095238095, 0.5476190476190477, 0.5952380952380952, 0.5238095238095238, 0.5595238095238095, 0.5714285714285714, 0.5238095238095238, 0.5714285714285714, 0.5952380952380952]\n",
    "conrem_acu_le6chars_1 = [0.6071428571428571, 0.6428571428571429, 0.5952380952380952, 0.6190476190476191, 0.6071428571428571, 0.5595238095238095, 0.5714285714285714, 0.6547619047619048, 0.6190476190476191, 0.6309523809523809]\n",
    "conrem_acu_le7chars_1 = [0.5833333333333334, 0.6190476190476191, 0.6071428571428571, 0.6666666666666666, 0.6071428571428571, 0.6190476190476191, 0.6071428571428571, 0.6428571428571429, 0.6190476190476191, 0.6071428571428571]\n",
    "conrem_acu_le8chars_1 = [0.5833333333333334, 0.6190476190476191, 0.6428571428571429, 0.6785714285714286, 0.7023809523809523, 0.6666666666666666, 0.6190476190476191, 0.5952380952380952, 0.5952380952380952, 0.6190476190476191]\n",
    "conrem_acu_le9chars_1 = [0.5595238095238095, 0.5952380952380952, 0.5714285714285714, 0.5595238095238095, 0.6428571428571429, 0.5714285714285714, 0.5952380952380952, 0.5714285714285714, 0.6547619047619048, 0.6190476190476191]\n",
    "conrem_acu_le10chars_1 = [0.5833333333333334, 0.5833333333333334, 0.5952380952380952, 0.6190476190476191, 0.6428571428571429, 0.6190476190476191, 0.6190476190476191, 0.5833333333333334, 0.5833333333333334, 0.5714285714285714]\n",
    "conrem_acu_le11chars_1 = [0.5952380952380952, 0.6309523809523809, 0.6071428571428571, 0.5476190476190477, 0.6428571428571429, 0.5833333333333334, 0.6309523809523809, 0.6190476190476191, 0.6428571428571429, 0.5833333333333334]\n",
    "conrem_acu_le12chars_1 = [0.5952380952380952, 0.5952380952380952, 0.5833333333333334, 0.5952380952380952, 0.5833333333333334, 0.6309523809523809, 0.6071428571428571, 0.5714285714285714, 0.6309523809523809, 0.5595238095238095]\n",
    "conrem_acu_le13chars_1 = [0.5833333333333334, 0.5833333333333334, 0.6190476190476191, 0.5833333333333334, 0.5952380952380952, 0.6190476190476191, 0.6428571428571429, 0.6190476190476191, 0.6190476190476191, 0.5714285714285714]\n",
    "conrem_acu_le14chars_1 = [0.6190476190476191, 0.6309523809523809, 0.5952380952380952, 0.6428571428571429, 0.5595238095238095, 0.5595238095238095, 0.5833333333333334, 0.5714285714285714, 0.5952380952380952, 0.5357142857142857]\n",
    "conrem_acu_le15chars_1 = [0.6071428571428571, 0.5952380952380952, 0.6190476190476191, 0.5952380952380952, 0.6190476190476191, 0.6071428571428571, 0.6190476190476191, 0.6071428571428571, 0.6190476190476191, 0.6190476190476191]\n",
    "conrem_acu_le16chars_1 = [0.6309523809523809, 0.6309523809523809, 0.5952380952380952, 0.5833333333333334, 0.6071428571428571, 0.5833333333333334, 0.6071428571428571, 0.5952380952380952, 0.6190476190476191, 0.6428571428571429]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "le1words_2 = [0.3291139240506329, 0.35443037974683544, 0.2911392405063291, 0.2911392405063291, 0.27848101265822783, 0.34177215189873417, 0.34177215189873417, 0.3670886075949367, 0.34177215189873417, 0.3924050632911392]\n",
    "le2words_2 = [0.5063291139240507, 0.45569620253164556, 0.3670886075949367, 0.3670886075949367, 0.43037974683544306, 0.3924050632911392, 0.4936708860759494, 0.4177215189873418, 0.4177215189873418, 0.4936708860759494]\n",
    "le3words_2 = [0.45569620253164556, 0.5189873417721519, 0.43037974683544306, 0.4050632911392405, 0.4430379746835443, 0.45569620253164556, 0.3924050632911392, 0.4936708860759494, 0.43037974683544306, 0.35443037974683544]\n",
    "le3chars_2 = [0.5063291139240507, 0.5063291139240507, 0.4177215189873418, 0.4810126582278481, 0.4810126582278481, 0.379746835443038, 0.45569620253164556, 0.5316455696202531, 0.34177215189873417, 0.43037974683544306]\n",
    "le4chars_2 = [0.43037974683544306, 0.35443037974683544, 0.4430379746835443, 0.46835443037974683, 0.5063291139240507, 0.4177215189873418, 0.46835443037974683, 0.4430379746835443, 0.43037974683544306, 0.3924050632911392]\n",
    "le5chars_2 = [0.3924050632911392, 0.3924050632911392, 0.4050632911392405, 0.3670886075949367, 0.4177215189873418, 0.379746835443038, 0.3670886075949367, 0.45569620253164556, 0.34177215189873417, 0.35443037974683544]\n",
    "le6chars_2 = [0.3924050632911392, 0.569620253164557, 0.4810126582278481, 0.5569620253164557, 0.43037974683544306, 0.46835443037974683, 0.45569620253164556, 0.43037974683544306, 0.45569620253164556, 0.4936708860759494]\n",
    "le7chars_2 = [0.46835443037974683, 0.4430379746835443, 0.4430379746835443, 0.4936708860759494, 0.46835443037974683, 0.4936708860759494, 0.4936708860759494, 0.5063291139240507, 0.45569620253164556, 0.5063291139240507]\n",
    "le8chars_2 = [0.46835443037974683, 0.46835443037974683, 0.4177215189873418, 0.43037974683544306, 0.46835443037974683, 0.45569620253164556, 0.4050632911392405, 0.4177215189873418, 0.45569620253164556, 0.4050632911392405]\n",
    "le9chars_2 = [0.43037974683544306, 0.46835443037974683, 0.45569620253164556, 0.45569620253164556, 0.4810126582278481, 0.3670886075949367, 0.45569620253164556, 0.43037974683544306, 0.4050632911392405, 0.45569620253164556]\n",
    "le10chars_2 = [0.4430379746835443, 0.3924050632911392, 0.4430379746835443, 0.4810126582278481, 0.4050632911392405, 0.4177215189873418, 0.4430379746835443, 0.4177215189873418, 0.3924050632911392, 0.4177215189873418]\n",
    "le11chars_2 = [0.35443037974683544, 0.3291139240506329, 0.43037974683544306, 0.4050632911392405, 0.4050632911392405, 0.3924050632911392, 0.4050632911392405, 0.3924050632911392, 0.43037974683544306, 0.35443037974683544]\n",
    "le12chars_2 = [0.35443037974683544, 0.3670886075949367, 0.31645569620253167, 0.4810126582278481, 0.35443037974683544, 0.379746835443038, 0.35443037974683544, 0.34177215189873417, 0.43037974683544306, 0.4430379746835443]\n",
    "le13chars_2 = [0.3291139240506329, 0.3670886075949367, 0.25316455696202533, 0.2911392405063291, 0.31645569620253167, 0.3670886075949367, 0.25316455696202533, 0.34177215189873417, 0.34177215189873417, 0.3037974683544304]\n",
    "le14chars_2 = [0.31645569620253167, 0.3037974683544304, 0.3037974683544304, 0.35443037974683544, 0.25316455696202533, 0.31645569620253167, 0.27848101265822783, 0.27848101265822783, 0.379746835443038, 0.22784810126582278]\n",
    "le15chars_2 = [0.3924050632911392, 0.35443037974683544, 0.27848101265822783, 0.34177215189873417, 0.2911392405063291, 0.3291139240506329, 0.3037974683544304, 0.3037974683544304, 0.3291139240506329, 0.3670886075949367]\n",
    "le16chars_2 = [0.34177215189873417, 0.25316455696202533, 0.4050632911392405, 0.3670886075949367, 0.35443037974683544, 0.35443037974683544, 0.3291139240506329, 0.3291139240506329, 0.379746835443038, 0.3670886075949367]\n",
    "ac1words_2 = [0.4430379746835443, 0.4050632911392405, 0.45569620253164556, 0.3924050632911392, 0.4810126582278481, 0.3291139240506329, 0.4177215189873418, 0.43037974683544306, 0.31645569620253167, 0.4810126582278481]\n",
    "ac2words_2 = [0.379746835443038, 0.4936708860759494, 0.5063291139240507, 0.4810126582278481, 0.3670886075949367, 0.43037974683544306, 0.4430379746835443, 0.46835443037974683, 0.4936708860759494, 0.4810126582278481]\n",
    "ac3words_2 = [0.43037974683544306, 0.5189873417721519, 0.4810126582278481, 0.4177215189873418, 0.45569620253164556, 0.46835443037974683, 0.4936708860759494, 0.4177215189873418, 0.4430379746835443, 0.5189873417721519]\n",
    "ac3chars_2 = [0.45569620253164556, 0.4430379746835443, 0.4810126582278481, 0.45569620253164556, 0.5189873417721519, 0.5189873417721519, 0.4810126582278481, 0.5569620253164557, 0.5443037974683544, 0.569620253164557]\n",
    "ac4chars_2 = [0.4936708860759494, 0.5443037974683544, 0.4810126582278481, 0.4430379746835443, 0.5569620253164557, 0.4810126582278481, 0.4810126582278481, 0.5569620253164557, 0.5569620253164557, 0.45569620253164556]\n",
    "ac5chars_2 = [0.4430379746835443, 0.43037974683544306, 0.4936708860759494, 0.43037974683544306, 0.45569620253164556, 0.5189873417721519, 0.3924050632911392, 0.45569620253164556, 0.4810126582278481, 0.4177215189873418]\n",
    "ac6chars_2 = [0.5189873417721519, 0.5316455696202531, 0.5569620253164557, 0.46835443037974683, 0.5316455696202531, 0.5063291139240507, 0.4810126582278481, 0.5316455696202531, 0.4936708860759494, 0.4430379746835443]\n",
    "ac7chars_2 = [0.4810126582278481, 0.5189873417721519, 0.4430379746835443, 0.4810126582278481, 0.4810126582278481, 0.5063291139240507, 0.46835443037974683, 0.5316455696202531, 0.5822784810126582, 0.5949367088607594]\n",
    "ac8chars_2 = [0.5569620253164557, 0.5063291139240507, 0.4810126582278481, 0.4936708860759494, 0.5063291139240507, 0.5443037974683544, 0.4810126582278481, 0.4430379746835443, 0.5063291139240507, 0.569620253164557]\n",
    "ac9chars_2 = [0.45569620253164556, 0.5063291139240507, 0.46835443037974683, 0.43037974683544306, 0.5316455696202531, 0.45569620253164556, 0.5063291139240507, 0.5443037974683544, 0.5189873417721519, 0.5063291139240507]\n",
    "ac10chars_2 = [0.5063291139240507, 0.4050632911392405, 0.4810126582278481, 0.43037974683544306, 0.5569620253164557, 0.4430379746835443, 0.4810126582278481, 0.4936708860759494, 0.5063291139240507, 0.5569620253164557]\n",
    "ac11chars_2 = [0.4430379746835443, 0.3924050632911392, 0.46835443037974683, 0.4936708860759494, 0.4050632911392405, 0.4430379746835443, 0.43037974683544306, 0.4177215189873418, 0.45569620253164556, 0.5316455696202531]\n",
    "ac12chars_2 = [0.43037974683544306, 0.379746835443038, 0.3291139240506329, 0.45569620253164556, 0.43037974683544306, 0.45569620253164556, 0.46835443037974683, 0.4050632911392405, 0.43037974683544306, 0.4050632911392405]\n",
    "ac13chars_2 = [0.379746835443038, 0.45569620253164556, 0.4430379746835443, 0.43037974683544306, 0.4050632911392405, 0.4810126582278481, 0.43037974683544306, 0.4810126582278481, 0.46835443037974683, 0.43037974683544306]\n",
    "ac14chars_2 = [0.379746835443038, 0.43037974683544306, 0.379746835443038, 0.46835443037974683, 0.31645569620253167, 0.4430379746835443, 0.3924050632911392, 0.4430379746835443, 0.379746835443038, 0.4050632911392405]\n",
    "ac15chars_2 = [0.46835443037974683, 0.43037974683544306, 0.4430379746835443, 0.34177215189873417, 0.3924050632911392, 0.43037974683544306, 0.4050632911392405, 0.4810126582278481, 0.34177215189873417, 0.3670886075949367]\n",
    "ac16chars_2 = [0.4936708860759494, 0.3924050632911392, 0.4430379746835443, 0.4177215189873418, 0.4050632911392405, 0.4050632911392405, 0.4177215189873418, 0.35443037974683544, 0.379746835443038, 0.43037974683544306]\n",
    "\n",
    "con_acurem_le1words_2 = [0.5316455696202531, 0.5822784810126582, 0.6455696202531646, 0.5443037974683544, 0.5949367088607594, 0.620253164556962, 0.5949367088607594, 0.6455696202531646, 0.5949367088607594, 0.569620253164557]\n",
    "con_acurem_le2words_2 = [0.5316455696202531, 0.5063291139240507, 0.45569620253164556, 0.45569620253164556, 0.5822784810126582, 0.5569620253164557, 0.5316455696202531, 0.5189873417721519, 0.46835443037974683, 0.4430379746835443]\n",
    "con_acurem_le3words_2 = [0.6329113924050633, 0.6329113924050633, 0.6582278481012658, 0.6962025316455697, 0.6835443037974683, 0.6708860759493671, 0.6455696202531646, 0.6455696202531646, 0.7088607594936709, 0.7088607594936709]\n",
    "con_acurem_le3chars_2 = [0.6329113924050633, 0.6455696202531646, 0.6708860759493671, 0.5443037974683544, 0.620253164556962, 0.6329113924050633, 0.5569620253164557, 0.5569620253164557, 0.569620253164557, 0.5949367088607594]\n",
    "con_acurem_le4chars_2 = [0.620253164556962, 0.6075949367088608, 0.5949367088607594, 0.6329113924050633, 0.6329113924050633, 0.6708860759493671, 0.6455696202531646, 0.6329113924050633, 0.6075949367088608, 0.6582278481012658]\n",
    "con_acurem_le5chars_2 = [0.6835443037974683, 0.620253164556962, 0.6075949367088608, 0.6835443037974683, 0.6582278481012658, 0.6708860759493671, 0.6329113924050633, 0.6455696202531646, 0.6329113924050633, 0.6582278481012658]\n",
    "con_acurem_le6chars_2 = [0.5949367088607594, 0.6329113924050633, 0.5822784810126582, 0.6075949367088608, 0.5949367088607594, 0.5822784810126582, 0.5569620253164557, 0.6455696202531646, 0.5822784810126582, 0.5569620253164557]\n",
    "con_acurem_le7chars_2 = [0.6329113924050633, 0.6075949367088608, 0.6075949367088608, 0.620253164556962, 0.5822784810126582, 0.6075949367088608, 0.6075949367088608, 0.5949367088607594, 0.5949367088607594, 0.620253164556962]\n",
    "con_acurem_le8chars_2 = [0.620253164556962, 0.620253164556962, 0.6455696202531646, 0.6455696202531646, 0.620253164556962, 0.6075949367088608, 0.6075949367088608, 0.6962025316455697, 0.620253164556962, 0.620253164556962]\n",
    "con_acurem_le9chars_2 = [0.6835443037974683, 0.6075949367088608, 0.6582278481012658, 0.6582278481012658, 0.7215189873417721, 0.6075949367088608, 0.6582278481012658, 0.7088607594936709, 0.6835443037974683, 0.6329113924050633]\n",
    "con_acurem_le10chars_2 = [0.6708860759493671, 0.5949367088607594, 0.6455696202531646, 0.6582278481012658, 0.620253164556962, 0.6329113924050633, 0.7215189873417721, 0.5949367088607594, 0.620253164556962, 0.6962025316455697]\n",
    "con_acurem_le11chars_2 = [0.6455696202531646, 0.6582278481012658, 0.7341772151898734, 0.6582278481012658, 0.6708860759493671, 0.6329113924050633, 0.6329113924050633, 0.6708860759493671, 0.7088607594936709, 0.6708860759493671]\n",
    "con_acurem_le12chars_2 = [0.6708860759493671, 0.6582278481012658, 0.6708860759493671, 0.5569620253164557, 0.6455696202531646, 0.6835443037974683, 0.6455696202531646, 0.6455696202531646, 0.6329113924050633, 0.6455696202531646]\n",
    "con_acurem_le13chars_2 = [0.5822784810126582, 0.569620253164557, 0.6455696202531646, 0.569620253164557, 0.620253164556962, 0.6455696202531646, 0.569620253164557, 0.6708860759493671, 0.6455696202531646, 0.5949367088607594]\n",
    "con_acurem_le14chars_2 = [0.5949367088607594, 0.5949367088607594, 0.620253164556962, 0.6075949367088608, 0.5822784810126582, 0.6075949367088608, 0.5822784810126582, 0.5569620253164557, 0.5569620253164557, 0.569620253164557]\n",
    "con_acurem_le15chars_2 = [0.5443037974683544, 0.5316455696202531, 0.5569620253164557, 0.5316455696202531, 0.5949367088607594, 0.569620253164557, 0.5316455696202531, 0.5316455696202531, 0.5189873417721519, 0.569620253164557]\n",
    "con_acurem_le16chars_2 = [0.5443037974683544, 0.5316455696202531, 0.5443037974683544, 0.5569620253164557, 0.5822784810126582, 0.569620253164557, 0.5443037974683544, 0.5189873417721519, 0.5569620253164557, 0.5189873417721519]\n",
    "con_acurem_ac1words_2 = [0.5316455696202531, 0.5443037974683544, 0.5949367088607594, 0.5189873417721519, 0.569620253164557, 0.6329113924050633, 0.4936708860759494, 0.5822784810126582, 0.569620253164557, 0.5569620253164557]\n",
    "con_acurem_ac2words_2 = [0.5443037974683544, 0.5822784810126582, 0.46835443037974683, 0.43037974683544306, 0.4810126582278481, 0.5443037974683544, 0.5443037974683544, 0.5063291139240507, 0.5443037974683544, 0.5063291139240507]\n",
    "con_acurem_ac3words_2 = [0.569620253164557, 0.6708860759493671, 0.6455696202531646, 0.6455696202531646, 0.6329113924050633, 0.620253164556962, 0.5569620253164557, 0.6708860759493671, 0.6075949367088608, 0.5822784810126582]\n",
    "con_acurem_ac3chars_2 = [0.6708860759493671, 0.620253164556962, 0.6075949367088608, 0.6329113924050633, 0.5949367088607594, 0.6455696202531646, 0.569620253164557, 0.6075949367088608, 0.6582278481012658, 0.6329113924050633]\n",
    "con_acurem_ac4chars_2 = [0.6329113924050633, 0.620253164556962, 0.6329113924050633, 0.6075949367088608, 0.6455696202531646, 0.6455696202531646, 0.5822784810126582, 0.6329113924050633, 0.620253164556962, 0.6329113924050633]\n",
    "con_acurem_ac5chars_2 = [0.5949367088607594, 0.6075949367088608, 0.6329113924050633, 0.6455696202531646, 0.5949367088607594, 0.6329113924050633, 0.6329113924050633, 0.5949367088607594, 0.5949367088607594, 0.620253164556962]\n",
    "con_acurem_ac6chars_2 = [0.5822784810126582, 0.6075949367088608, 0.6329113924050633, 0.5569620253164557, 0.6329113924050633, 0.6329113924050633, 0.6708860759493671, 0.620253164556962, 0.6329113924050633, 0.620253164556962]\n",
    "con_acurem_ac7chars_2 = [0.5443037974683544, 0.5822784810126582, 0.5822784810126582, 0.5949367088607594, 0.5949367088607594, 0.5822784810126582, 0.6075949367088608, 0.5822784810126582, 0.6455696202531646, 0.5569620253164557]\n",
    "con_acurem_ac8chars_2 = [0.620253164556962, 0.5569620253164557, 0.5949367088607594, 0.569620253164557, 0.5569620253164557, 0.6075949367088608, 0.5569620253164557, 0.5569620253164557, 0.620253164556962, 0.6075949367088608]\n",
    "con_acurem_ac9chars_2 = [0.569620253164557, 0.6455696202531646, 0.6075949367088608, 0.5949367088607594, 0.620253164556962, 0.6075949367088608, 0.6329113924050633, 0.620253164556962, 0.6075949367088608, 0.569620253164557]\n",
    "con_acurem_ac10chars_2 = [0.569620253164557, 0.6455696202531646, 0.6329113924050633, 0.5822784810126582, 0.5443037974683544, 0.6075949367088608, 0.5822784810126582, 0.5443037974683544, 0.620253164556962, 0.6329113924050633]\n",
    "con_acurem_ac11chars_2 = [0.6582278481012658, 0.6075949367088608, 0.6075949367088608, 0.6455696202531646, 0.6582278481012658, 0.6455696202531646, 0.620253164556962, 0.6075949367088608, 0.620253164556962, 0.6075949367088608]\n",
    "con_acurem_ac12chars_2 = [0.6582278481012658, 0.620253164556962, 0.6582278481012658, 0.5949367088607594, 0.6075949367088608, 0.5822784810126582, 0.620253164556962, 0.6075949367088608, 0.5822784810126582, 0.6329113924050633]\n",
    "con_acurem_ac13chars_2 = [0.569620253164557, 0.5569620253164557, 0.5569620253164557, 0.5822784810126582, 0.569620253164557, 0.5949367088607594, 0.5569620253164557, 0.6075949367088608, 0.620253164556962, 0.569620253164557]\n",
    "con_acurem_ac14chars_2 = [0.569620253164557, 0.569620253164557, 0.5443037974683544, 0.5443037974683544, 0.5443037974683544, 0.5569620253164557, 0.569620253164557, 0.569620253164557, 0.5822784810126582, 0.569620253164557]\n",
    "con_acurem_ac15chars_2 = [0.5822784810126582, 0.5316455696202531, 0.5316455696202531, 0.5569620253164557, 0.5189873417721519, 0.5443037974683544, 0.569620253164557, 0.5443037974683544, 0.5443037974683544, 0.5443037974683544]\n",
    "con_acurem_ac16chars_2 = [0.5063291139240507, 0.5316455696202531, 0.45569620253164556, 0.5443037974683544, 0.4936708860759494, 0.5316455696202531, 0.5316455696202531, 0.5569620253164557, 0.5316455696202531, 0.5569620253164557]\n",
    "\n",
    "\n",
    "conrem_acu_ac1words_2 = [0.7341772151898734, 0.6708860759493671, 0.7088607594936709, 0.6962025316455697, 0.7341772151898734, 0.6962025316455697, 0.6329113924050633, 0.6708860759493671, 0.7468354430379747, 0.6835443037974683]\n",
    "conrem_acu_ac2words_2 = [0.7974683544303798, 0.7974683544303798, 0.810126582278481, 0.8227848101265823, 0.7848101265822784, 0.7848101265822784, 0.7974683544303798, 0.7341772151898734, 0.810126582278481, 0.7721518987341772]\n",
    "conrem_acu_ac3words_2 = [0.7468354430379747, 0.7848101265822784, 0.759493670886076, 0.7721518987341772, 0.6962025316455697, 0.7721518987341772, 0.759493670886076, 0.7215189873417721, 0.7341772151898734, 0.7721518987341772]\n",
    "conrem_acu_ac3chars_2 = [0.7468354430379747, 0.8354430379746836, 0.810126582278481, 0.8354430379746836, 0.7215189873417721, 0.7721518987341772, 0.7721518987341772, 0.7974683544303798, 0.7215189873417721, 0.7974683544303798]\n",
    "conrem_acu_ac4chars_2 = [0.7088607594936709, 0.759493670886076, 0.7341772151898734, 0.6455696202531646, 0.7088607594936709, 0.7215189873417721, 0.7468354430379747, 0.7848101265822784, 0.7848101265822784, 0.7215189873417721]\n",
    "conrem_acu_ac5chars_2 = [0.7468354430379747, 0.6962025316455697, 0.8354430379746836, 0.7215189873417721, 0.7721518987341772, 0.759493670886076, 0.7468354430379747, 0.7468354430379747, 0.7088607594936709, 0.7721518987341772]\n",
    "conrem_acu_ac6chars_2 = [0.8607594936708861, 0.8227848101265823, 0.7974683544303798, 0.7848101265822784, 0.7848101265822784, 0.8481012658227848, 0.7848101265822784, 0.810126582278481, 0.8227848101265823, 0.7341772151898734]\n",
    "conrem_acu_ac7chars_2 = [0.810126582278481, 0.810126582278481, 0.8227848101265823, 0.7848101265822784, 0.8734177215189873, 0.810126582278481, 0.7974683544303798, 0.7974683544303798, 0.810126582278481, 0.8481012658227848]\n",
    "conrem_acu_ac8chars_2 = [0.7974683544303798, 0.7721518987341772, 0.7974683544303798, 0.7974683544303798, 0.810126582278481, 0.7974683544303798, 0.7721518987341772, 0.8227848101265823, 0.7848101265822784, 0.7848101265822784]\n",
    "conrem_acu_ac9chars_2 = [0.7341772151898734, 0.7848101265822784, 0.7848101265822784, 0.7721518987341772, 0.759493670886076, 0.7468354430379747, 0.7468354430379747, 0.7848101265822784, 0.7721518987341772, 0.7721518987341772]\n",
    "conrem_acu_ac10chars_2 = [0.6962025316455697, 0.7088607594936709, 0.7215189873417721, 0.6962025316455697, 0.7721518987341772, 0.7341772151898734, 0.7088607594936709, 0.7215189873417721, 0.759493670886076, 0.7468354430379747]\n",
    "conrem_acu_ac11chars_2 = [0.6962025316455697, 0.7088607594936709, 0.7215189873417721, 0.7215189873417721, 0.7215189873417721, 0.7215189873417721, 0.6835443037974683, 0.6962025316455697, 0.7215189873417721, 0.7215189873417721]\n",
    "conrem_acu_ac12chars_2 = [0.7215189873417721, 0.6835443037974683, 0.6708860759493671, 0.7341772151898734, 0.7088607594936709, 0.6708860759493671, 0.6962025316455697, 0.7215189873417721, 0.6582278481012658, 0.6708860759493671]\n",
    "conrem_acu_ac13chars_2 = [0.6835443037974683, 0.7341772151898734, 0.6835443037974683, 0.6708860759493671, 0.7341772151898734, 0.7468354430379747, 0.6708860759493671, 0.7468354430379747, 0.6835443037974683, 0.7341772151898734]\n",
    "conrem_acu_ac14chars_2 = [0.6962025316455697, 0.7088607594936709, 0.6455696202531646, 0.6835443037974683, 0.6835443037974683, 0.6708860759493671, 0.6708860759493671, 0.6708860759493671, 0.6962025316455697, 0.7215189873417721]\n",
    "conrem_acu_ac15chars_2 = [0.6582278481012658, 0.6455696202531646, 0.6962025316455697, 0.6708860759493671, 0.6329113924050633, 0.6835443037974683, 0.6582278481012658, 0.6962025316455697, 0.6835443037974683, 0.6708860759493671]\n",
    "conrem_acu_ac16chars_2 =[0.6708860759493671, 0.6455696202531646, 0.6708860759493671, 0.6329113924050633, 0.6455696202531646, 0.6582278481012658, 0.6962025316455697, 0.6962025316455697, 0.6835443037974683, 0.6329113924050633]\n",
    "conrem_acu_le1words_2 = [0.569620253164557, 0.5949367088607594, 0.5189873417721519, 0.43037974683544306, 0.43037974683544306, 0.4936708860759494, 0.5316455696202531, 0.5822784810126582, 0.620253164556962, 0.569620253164557]\n",
    "conrem_acu_le2words_2 = [0.7341772151898734, 0.6962025316455697, 0.759493670886076, 0.6455696202531646, 0.7468354430379747, 0.7215189873417721, 0.7341772151898734, 0.7088607594936709, 0.7215189873417721, 0.6455696202531646]\n",
    "conrem_acu_le3words_2 = [0.7088607594936709, 0.6455696202531646, 0.6455696202531646, 0.6962025316455697, 0.6835443037974683, 0.6835443037974683, 0.759493670886076, 0.620253164556962, 0.7088607594936709, 0.759493670886076]\n",
    "conrem_acu_le3chars_2 = [0.7341772151898734, 0.6708860759493671, 0.6962025316455697, 0.6962025316455697, 0.569620253164557, 0.6582278481012658, 0.7215189873417721, 0.6835443037974683, 0.6835443037974683, 0.7215189873417721]\n",
    "conrem_acu_le4chars_2 = [0.620253164556962, 0.6708860759493671, 0.6582278481012658, 0.6708860759493671, 0.5949367088607594, 0.6962025316455697, 0.6962025316455697, 0.620253164556962, 0.6455696202531646, 0.6455696202531646]\n",
    "conrem_acu_le5chars_2 = [0.6962025316455697, 0.6835443037974683, 0.6455696202531646, 0.6708860759493671, 0.6962025316455697, 0.6708860759493671, 0.5822784810126582, 0.6835443037974683, 0.6835443037974683, 0.6329113924050633]\n",
    "conrem_acu_le6chars_2 = [0.6962025316455697, 0.7341772151898734, 0.6962025316455697, 0.6962025316455697, 0.6962025316455697, 0.6455696202531646, 0.6962025316455697, 0.6455696202531646, 0.7215189873417721, 0.759493670886076]\n",
    "conrem_acu_le7chars_2 = [0.6962025316455697, 0.6835443037974683, 0.7215189873417721, 0.7341772151898734, 0.6835443037974683, 0.6962025316455697, 0.6708860759493671, 0.6962025316455697, 0.6708860759493671, 0.6708860759493671]\n",
    "conrem_acu_le8chars_2 = [0.6835443037974683, 0.6582278481012658, 0.6582278481012658, 0.6962025316455697, 0.7088607594936709, 0.620253164556962, 0.6582278481012658, 0.7215189873417721, 0.7215189873417721, 0.6708860759493671]\n",
    "conrem_acu_le9chars_2 = [0.6582278481012658, 0.6455696202531646, 0.6582278481012658, 0.6708860759493671, 0.6708860759493671, 0.620253164556962, 0.6075949367088608, 0.6582278481012658, 0.620253164556962, 0.6582278481012658]\n",
    "conrem_acu_le10chars_2 = [0.620253164556962, 0.6708860759493671, 0.6075949367088608, 0.6835443037974683, 0.5822784810126582, 0.6329113924050633, 0.6455696202531646, 0.620253164556962, 0.6708860759493671, 0.5316455696202531]\n",
    "conrem_acu_le11chars_2 = [0.6329113924050633, 0.5822784810126582, 0.6329113924050633, 0.6329113924050633, 0.569620253164557, 0.6075949367088608, 0.5822784810126582, 0.6075949367088608, 0.6075949367088608, 0.5822784810126582]\n",
    "conrem_acu_le12chars_2 = [0.620253164556962, 0.5443037974683544, 0.5822784810126582, 0.5949367088607594, 0.569620253164557, 0.5949367088607594, 0.5949367088607594, 0.5949367088607594, 0.5822784810126582, 0.6075949367088608]\n",
    "conrem_acu_le13chars_2 = [0.569620253164557, 0.5569620253164557, 0.5189873417721519, 0.5443037974683544, 0.5569620253164557, 0.5189873417721519, 0.569620253164557, 0.6075949367088608, 0.569620253164557, 0.5443037974683544]\n",
    "conrem_acu_le14chars_2 = [0.6075949367088608, 0.5822784810126582, 0.5569620253164557, 0.5443037974683544, 0.5063291139240507, 0.569620253164557, 0.5443037974683544, 0.5316455696202531, 0.5569620253164557, 0.5316455696202531]\n",
    "conrem_acu_le15chars_2 = [0.5822784810126582, 0.5949367088607594, 0.569620253164557, 0.5443037974683544, 0.6075949367088608, 0.5822784810126582, 0.5569620253164557, 0.6329113924050633, 0.5569620253164557, 0.6075949367088608]\n",
    "conrem_acu_le16chars_2 = [0.569620253164557, 0.5822784810126582, 0.5443037974683544, 0.5063291139240507, 0.5949367088607594, 0.5569620253164557, 0.620253164556962, 0.5316455696202531, 0.5822784810126582, 0.5443037974683544]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Taaloefening 1+2, three groups\n",
    "le1words = [0.38650306748466257, 0.2883435582822086, 0.32515337423312884, 0.26993865030674846, 0.3312883435582822, 0.27607361963190186, 0.2883435582822086, 0.3006134969325153, 0.294478527607362, 0.36809815950920244]\n",
    "le2words = [0.31901840490797545, 0.36809815950920244, 0.3619631901840491, 0.36809815950920244, 0.37423312883435583, 0.3374233128834356, 0.27607361963190186, 0.3558282208588957, 0.37423312883435583, 0.31901840490797545]\n",
    "le3words = [0.4601226993865031, 0.4294478527607362, 0.44171779141104295, 0.38650306748466257, 0.4110429447852761, 0.4294478527607362, 0.4110429447852761, 0.37423312883435583, 0.4294478527607362, 0.4662576687116564]\n",
    "le3chars = [0.44171779141104295, 0.4785276073619632, 0.43558282208588955, 0.5276073619631901, 0.39263803680981596, 0.4785276073619632, 0.4049079754601227, 0.4294478527607362, 0.4539877300613497, 0.4233128834355828]\n",
    "le4chars = [0.43558282208588955, 0.4233128834355828, 0.44785276073619634, 0.4049079754601227, 0.3558282208588957, 0.4294478527607362, 0.44785276073619634, 0.3374233128834356, 0.4171779141104294]\n",
    "le5chars = [0.3496932515337423, 0.37423312883435583, 0.3374233128834356, 0.31901840490797545, 0.3312883435582822, 0.3619631901840491, 0.36809815950920244, 0.3374233128834356, 0.38650306748466257, 0.3496932515337423]\n",
    "le6chars = [0.3803680981595092, 0.3374233128834356, 0.3496932515337423, 0.3558282208588957, 0.3374233128834356, 0.3496932515337423, 0.3558282208588957, 0.3312883435582822, 0.3312883435582822, 0.3619631901840491]\n",
    "le7chars = [0.34355828220858897, 0.3803680981595092, 0.3619631901840491, 0.3312883435582822, 0.36809815950920244, 0.3558282208588957, 0.3619631901840491, 0.4171779141104294, 0.3006134969325153, 0.3374233128834356]\n",
    "le8chars = [0.3558282208588957, 0.34355828220858897, 0.37423312883435583, 0.34355828220858897, 0.4049079754601227, 0.3067484662576687, 0.3374233128834356, 0.37423312883435583, 0.3374233128834356, 0.3006134969325153]\n",
    "le9chars = [0.31901840490797545, 0.27607361963190186, 0.3558282208588957, 0.3374233128834356, 0.3374233128834356, 0.32515337423312884, 0.3558282208588957, 0.3619631901840491, 0.31901840490797545, 0.3619631901840491]\n",
    "le10chars = [0.3496932515337423, 0.31901840490797545, 0.3374233128834356, 0.3128834355828221, 0.31901840490797545, 0.26993865030674846, 0.37423312883435583, 0.34355828220858897, 0.3558282208588957, 0.3558282208588957]\n",
    "le11chars = [0.2883435582822086, 0.3803680981595092, 0.37423312883435583, 0.34355828220858897, 0.3067484662576687, 0.32515337423312884, 0.3803680981595092, 0.3619631901840491, 0.32515337423312884, 0.3067484662576687]\n",
    "le12chars = [0.3312883435582822, 0.3803680981595092, 0.34355828220858897, 0.3374233128834356, 0.3006134969325153, 0.3496932515337423, 0.36809815950920244, 0.3312883435582822, 0.3312883435582822, 0.36809815950920244]\n",
    "le13chars = [0.3312883435582822, 0.3312883435582822, 0.31901840490797545, 0.3496932515337423, 0.3312883435582822, 0.3067484662576687, 0.3067484662576687, 0.3374233128834356, 0.36809815950920244, 0.3312883435582822]\n",
    "le14chars = [0.3312883435582822, 0.37423312883435583, 0.2822085889570552, 0.3496932515337423, 0.3803680981595092, 0.3067484662576687, 0.31901840490797545, 0.24539877300613497, 0.3496932515337423, 0.27607361963190186]\n",
    "le15chars = [0.3067484662576687, 0.3128834355828221, 0.3803680981595092, 0.3496932515337423, 0.3803680981595092, 0.38650306748466257, 0.3496932515337423, 0.3496932515337423, 0.3128834355828221, 0.3619631901840491]\n",
    "le16chars = [0.32515337423312884, 0.36809815950920244, 0.3312883435582822, 0.3987730061349693, 0.36809815950920244, 0.39263803680981596, 0.38650306748466257, 0.3558282208588957, 0.294478527607362, 0.3312883435582822]\n",
    "ac1words = [0.4233128834355828, 0.4539877300613497, 0.3987730061349693, 0.44171779141104295, 0.3987730061349693, 0.38650306748466257, 0.3987730061349693, 0.4171779141104294, 0.4049079754601227, 0.4110429447852761]\n",
    "ac2words = [0.37423312883435583, 0.3803680981595092, 0.3987730061349693, 0.32515337423312884, 0.37423312883435583, 0.38650306748466257, 0.3803680981595092, 0.3558282208588957, 0.32515337423312884, 0.3128834355828221]\n",
    "ac3words = [0.44171779141104295, 0.4601226993865031, 0.4294478527607362, 0.3987730061349693, 0.48466257668711654, 0.4049079754601227, 0.3987730061349693, 0.4601226993865031, 0.4601226993865031, 0.44171779141104295]\n",
    "ac8chars = [0.4049079754601227,0.39263803680981596, 0.3803680981595092, 0.3374233128834356, 0.39263803680981596, 0.3619631901840491, 0.3496932515337423, 0.3987730061349693, 0.3558282208588957, 0.3803680981595092]\n",
    "ac7chars = [0.39263803680981596, 0.37423312883435583, 0.3496932515337423, 0.4049079754601227, 0.4110429447852761,0.4049079754601227, 0.4171779141104294, 0.3558282208588957, 0.3803680981595092,0.37423312883435583]\n",
    "ac6chars = [0.3496932515337423,0.39263803680981596,0.4049079754601227,0.4171779141104294,0.3558282208588957,0.4049079754601227,0.39263803680981596,0.37423312883435583,0.44171779141104295,0.3619631901840491]\n",
    "ac5chars = [0.3987730061349693, 0.36809815950920244, 0.43558282208588955,0.37423312883435583, 0.4049079754601227, 0.4233128834355828, 0.4049079754601227,0.4049079754601227, 0.4171779141104294, 0.37423312883435583]\n",
    "ac4chars = [0.38650306748466257, 0.4662576687116564, 0.48466257668711654,0.4171779141104294, 0.39263803680981596, 0.44171779141104295, 0.4294478527607362,0.43558282208588955, 0.4233128834355828]\n",
    "ac3chars = [0.5398773006134969, 0.4539877300613497, 0.4294478527607362,0.49693251533742333, 0.48466257668711654, 0.49693251533742333, 0.4723926380368098,0.43558282208588955, 0.4233128834355828, 0.4171779141104294]\n",
    "ac9chars = [0.3558282208588957, 0.3987730061349693, 0.4049079754601227,0.4233128834355828, 0.36809815950920244, 0.39263803680981596,0.3312883435582822, 0.3374233128834356, 0.3803680981595092, 0.34355828220858897]\n",
    "ac10chars = [0.4233128834355828,0.3803680981595092, 0.3987730061349693, 0.38650306748466257, 0.3987730061349693, 0.3803680981595092, 0.39263803680981596, 0.4110429447852761, 0.4049079754601227, 0.3803680981595092]\n",
    "ac11chars = [0.3987730061349693, 0.37423312883435583, 0.4294478527607362, 0.43558282208588955, 0.4233128834355828, 0.4171779141104294, 0.4662576687116564, 0.4049079754601227, 0.39263803680981596, 0.4233128834355828]\n",
    "ac13chars = [0.32515337423312884, 0.3987730061349693, 0.36809815950920244, 0.4294478527607362, 0.3558282208588957, 0.4049079754601227, 0.4171779141104294, 0.36809815950920244, 0.3558282208588957, 0.4171779141104294]\n",
    "ac14chars = [0.3619631901840491, 0.3374233128834356, 0.4049079754601227, 0.3558282208588957, 0.3496932515337423, 0.37423312883435583, 0.3374233128834356, 0.38650306748466257, 0.37423312883435583, 0.3496932515337423]\n",
    "ac15chars = [0.34355828220858897, 0.4049079754601227, 0.3558282208588957, 0.3496932515337423, 0.36809815950920244, 0.4171779141104294, 0.37423312883435583, 0.3006134969325153, 0.3803680981595092, 0.3374233128834356]\n",
    "ac16chars = [0.36809815950920244, 0.3496932515337423, 0.43558282208588955, 0.34355828220858897, 0.37423312883435583, 0.3312883435582822, 0.4110429447852761, 0.3496932515337423,0.3558282208588957, 0.3374233128834356]\n",
    "# two groups\n",
    "con_acurem_le1words = [0.5521472392638037, 0.5276073619631901, 0.5214723926380368, 0.5644171779141104, 0.5521472392638037, 0.5644171779141104, 0.5398773006134969, 0.5153374233128835, 0.558282208588957, 0.5705521472392638]\n",
    "con_acurem_le2words = [0.5644171779141104, 0.5398773006134969, 0.588957055214724, 0.5337423312883436, 0.558282208588957, 0.50920245398773, 0.5337423312883436, 0.50920245398773, 0.5766871165644172, 0.5398773006134969]\n",
    "con_acurem_le3words = [0.588957055214724, 0.6073619631901841, 0.6319018404907976, 0.656441717791411, 0.6012269938650306, 0.6012269938650306, 0.6012269938650306, 0.6932515337423313, 0.5766871165644172, 0.6073619631901841]\n",
    "con_acurem_le3chars = [0.5521472392638037, 0.588957055214724, 0.5460122699386503, 0.6257668711656442, 0.5460122699386503, 0.5950920245398773, 0.5950920245398773, 0.5705521472392638, 0.5521472392638037, 0.5950920245398773]\n",
    "con_acurem_le4chars = [0.6380368098159509, 0.6687116564417178, 0.6073619631901841, 0.6380368098159509, 0.5950920245398773, 0.6196319018404908, 0.5766871165644172, 0.6134969325153374, 0.6134969325153374, 0.5766871165644172]\n",
    "con_acurem_le5chars = [0.6503067484662577, 0.6134969325153374, 0.588957055214724, 0.6257668711656442, 0.6319018404907976, 0.6012269938650306, 0.6134969325153374, 0.6319018404907976, 0.6134969325153374, 0.6012269938650306]\n",
    "con_acurem_le6chars = [0.6012269938650306, 0.588957055214724, 0.6073619631901841, 0.6073619631901841, 0.5766871165644172, 0.6012269938650306, 0.6257668711656442, 0.6134969325153374, 0.6441717791411042, 0.5950920245398773]\n",
    "con_acurem_le7chars = [0.6073619631901841, 0.588957055214724, 0.5950920245398773, 0.588957055214724, 0.6073619631901841, 0.6257668711656442, 0.6319018404907976, 0.6319018404907976, 0.6257668711656442, 0.6073619631901841]\n",
    "con_acurem_le8chars = [0.6319018404907976, 0.6319018404907976, 0.6319018404907976, 0.5950920245398773, 0.656441717791411, 0.6012269938650306, 0.6380368098159509, 0.656441717791411, 0.6134969325153374, 0.5950920245398773]\n",
    "con_acurem_le9chars = [0.6196319018404908, 0.6196319018404908, 0.5828220858895705, 0.5950920245398773, 0.5950920245398773, 0.6134969325153374, 0.5950920245398773, 0.5766871165644172, 0.6134969325153374, 0.6012269938650306]\n",
    "con_acurem_le10chars = [0.6012269938650306, 0.6196319018404908, 0.6012269938650306, 0.6134969325153374, 0.6319018404907976, 0.6134969325153374, 0.6012269938650306, 0.6073619631901841, 0.6196319018404908, 0.6196319018404908]\n",
    "con_acurem_le11chars = [0.6380368098159509, 0.6319018404907976, 0.6073619631901841, 0.6441717791411042, 0.5950920245398773, 0.6134969325153374, 0.6073619631901841, 0.6073619631901841, 0.5950920245398773, 0.6012269938650306]\n",
    "con_acurem_le12chars = [0.6012269938650306, 0.6134969325153374, 0.6319018404907976, 0.6134969325153374, 0.5828220858895705, 0.5950920245398773, 0.6625766871165644, 0.6073619631901841, 0.6134969325153374, 0.6073619631901841]\n",
    "con_acurem_le13chars = [0.5950920245398773, 0.5766871165644172, 0.6196319018404908, 0.6012269938650306, 0.6196319018404908, 0.6380368098159509, 0.5950920245398773, 0.6257668711656442, 0.5705521472392638, 0.6012269938650306]\n",
    "con_acurem_le14chars = [0.6073619631901841, 0.5950920245398773, 0.6134969325153374, 0.5828220858895705, 0.6134969325153374, 0.5644171779141104, 0.5705521472392638, 0.5828220858895705, 0.6257668711656442, 0.5950920245398773]\n",
    "con_acurem_le15chars = [0.5950920245398773, 0.5766871165644172, 0.6196319018404908, 0.558282208588957, 0.5828220858895705, 0.5828220858895705, 0.5705521472392638, 0.5644171779141104, 0.5705521472392638, 0.5828220858895705]\n",
    "con_acurem_le16chars = [0.5766871165644172, 0.5521472392638037, 0.588957055214724, 0.6073619631901841, 0.6134969325153374, 0.6073619631901841, 0.6012269938650306, 0.588957055214724, 0.6441717791411042, 0.6012269938650306]\n",
    "con_acurem_ac1words =[0.5398773006134969, 0.50920245398773, 0.588957055214724, 0.5398773006134969, 0.558282208588957, 0.5460122699386503, 0.558282208588957, 0.5521472392638037, 0.558282208588957, 0.6073619631901841]\n",
    "con_acurem_ac2words =[0.6012269938650306, 0.5644171779141104, 0.5950920245398773, 0.588957055214724, 0.5460122699386503, 0.5705521472392638, 0.588957055214724, 0.5521472392638037, 0.5705521472392638, 0.5521472392638037]\n",
    "con_acurem_ac3words =[0.588957055214724, 0.6134969325153374, 0.5828220858895705, 0.6503067484662577, 0.6012269938650306, 0.5644171779141104, 0.6073619631901841, 0.5828220858895705, 0.6196319018404908, 0.6196319018404908]\n",
    "con_acurem_ac4chars =[0.6625766871165644, 0.6257668711656442, 0.656441717791411, 0.6380368098159509, 0.588957055214724, 0.6012269938650306, 0.588957055214724, 0.6073619631901841, 0.6196319018404908, 0.6012269938650306]\n",
    "con_acurem_ac5chars =[0.6319018404907976, 0.656441717791411, 0.6196319018404908, 0.6380368098159509, 0.6257668711656442, 0.6257668711656442, 0.6380368098159509, 0.6503067484662577, 0.656441717791411, 0.6319018404907976]\n",
    "con_acurem_ac6chars =[0.6380368098159509, 0.6380368098159509, 0.6134969325153374, 0.6319018404907976, 0.6380368098159509, 0.6196319018404908, 0.6380368098159509, 0.656441717791411, 0.6503067484662577, 0.6257668711656442]\n",
    "con_acurem_ac7chars =[0.6012269938650306, 0.6134969325153374, 0.6196319018404908, 0.6134969325153374, 0.6257668711656442, 0.6257668711656442, 0.6380368098159509, 0.6319018404907976, 0.588957055214724, 0.6380368098159509]\n",
    "con_acurem_ac8chars =[0.6441717791411042, 0.6073619631901841, 0.6073619631901841, 0.6257668711656442, 0.6134969325153374, 0.6257668711656442, 0.6073619631901841, 0.6380368098159509, 0.6441717791411042, 0.6380368098159509]\n",
    "con_acurem_ac9chars =[0.6257668711656442, 0.6257668711656442, 0.588957055214724, 0.6134969325153374, 0.6012269938650306, 0.6380368098159509, 0.6134969325153374, 0.6503067484662577, 0.6503067484662577, 0.6319018404907976]\n",
    "con_acurem_ac10chars =[0.6380368098159509, 0.6380368098159509, 0.6012269938650306, 0.6134969325153374, 0.656441717791411, 0.6319018404907976, 0.6441717791411042, 0.6196319018404908, 0.6257668711656442, 0.5950920245398773]\n",
    "con_acurem_ac11chars =[0.6196319018404908, 0.6380368098159509, 0.6257668711656442, 0.6073619631901841, 0.6257668711656442, 0.6503067484662577, 0.6134969325153374, 0.6012269938650306, 0.6134969325153374, 0.6196319018404908]\n",
    "con_acurem_ac12chars =[0.5828220858895705, 0.6134969325153374, 0.6257668711656442, 0.6380368098159509, 0.6319018404907976, 0.6503067484662577, 0.6073619631901841, 0.5828220858895705, 0.6257668711656442, 0.6380368098159509]\n",
    "con_acurem_ac13chars =[0.6073619631901841, 0.6134969325153374, 0.6134969325153374, 0.6073619631901841, 0.6134969325153374, 0.6012269938650306, 0.5828220858895705, 0.5828220858895705, 0.6134969325153374, 0.6073619631901841]\n",
    "con_acurem_ac14chars =[0.6134969325153374, 0.6012269938650306, 0.6380368098159509, 0.6134969325153374, 0.588957055214724, 0.5828220858895705, 0.588957055214724, 0.558282208588957, 0.6012269938650306, 0.6196319018404908]\n",
    "con_acurem_ac15chars =[0.5950920245398773, 0.6012269938650306, 0.5950920245398773, 0.6073619631901841, 0.588957055214724, 0.6073619631901841, 0.5705521472392638, 0.5766871165644172, 0.5644171779141104, 0.5950920245398773]\n",
    "con_acurem_ac16chars =[0.588957055214724, 0.5950920245398773, 0.5705521472392638, 0.5766871165644172, 0.588957055214724, 0.5950920245398773, 0.5766871165644172, 0.5766871165644172, 0.5766871165644172, 0.6134969325153374]\n",
    "# two groups\n",
    "conrem_acu_le1words = [0.5460122699386503, 0.5950920245398773, 0.558282208588957, 0.5521472392638037, 0.5766871165644172, 0.5460122699386503, 0.5828220858895705, 0.5398773006134969, 0.5214723926380368, 0.50920245398773]\n",
    "conrem_acu_le2words = [0.6319018404907976, 0.6687116564417178, 0.6319018404907976, 0.6625766871165644, 0.6503067484662577, 0.5705521472392638, 0.6012269938650306, 0.6012269938650306, 0.6809815950920245, 0.6196319018404908]\n",
    "conrem_acu_le3words =[0.6073619631901841, 0.6319018404907976, 0.6748466257668712, 0.6012269938650306, 0.6012269938650306, 0.6625766871165644, 0.6073619631901841, 0.6441717791411042, 0.6319018404907976, 0.6257668711656442]\n",
    "conrem_acu_le3chars = [0.6073619631901841, 0.6012269938650306, 0.6257668711656442, 0.6380368098159509, 0.6441717791411042, 0.6257668711656442, 0.6380368098159509, 0.6196319018404908, 0.6380368098159509, 0.6748466257668712]\n",
    "conrem_acu_le4chars = [0.6134969325153374, 0.6012269938650306, 0.5644171779141104, 0.5950920245398773, 0.6012269938650306, 0.5766871165644172, 0.6257668711656442, 0.6257668711656442, 0.5766871165644172, 0.6012269938650306]\n",
    "conrem_acu_le5chars = [0.6196319018404908, 0.6073619631901841, 0.6134969325153374, 0.5950920245398773, 0.6319018404907976, 0.6073619631901841, 0.5766871165644172, 0.6012269938650306, 0.6134969325153374, 0.6196319018404908]\n",
    "conrem_acu_le6chars = [0.6380368098159509, 0.5950920245398773, 0.6748466257668712, 0.6625766871165644, 0.656441717791411, 0.656441717791411, 0.6257668711656442, 0.6134969325153374, 0.6687116564417178, 0.6625766871165644]\n",
    "conrem_acu_le7chars = [0.6748466257668712, 0.6380368098159509, 0.6503067484662577, 0.656441717791411, 0.6503067484662577, 0.6687116564417178, 0.6319018404907976, 0.6380368098159509, 0.6196319018404908, 0.6503067484662577]\n",
    "conrem_acu_le8chars = [0.6441717791411042, 0.6625766871165644, 0.6625766871165644, 0.6441717791411042, 0.6257668711656442, 0.6380368098159509, 0.6441717791411042, 0.6319018404907976, 0.6748466257668712, 0.6134969325153374]\n",
    "conrem_acu_le9chars = [0.6319018404907976, 0.6012269938650306, 0.6012269938650306, 0.6380368098159509, 0.6196319018404908, 0.6380368098159509, 0.6196319018404908, 0.6441717791411042, 0.6380368098159509, 0.6257668711656442]\n",
    "conrem_acu_le10chars = [0.6380368098159509, 0.6319018404907976, 0.6503067484662577, 0.6073619631901841, 0.6380368098159509, 0.5950920245398773, 0.6134969325153374, 0.6257668711656442, 0.6319018404907976, 0.6196319018404908]\n",
    "conrem_acu_le11chars = [0.588957055214724, 0.6380368098159509, 0.6319018404907976, 0.6441717791411042, 0.6073619631901841, 0.6257668711656442, 0.6380368098159509, 0.5950920245398773, 0.6073619631901841, 0.6073619631901841]\n",
    "conrem_acu_le12chars = [0.6441717791411042, 0.588957055214724, 0.5766871165644172, 0.5950920245398773, 0.6134969325153374, 0.6134969325153374, 0.5828220858895705, 0.5950920245398773, 0.5766871165644172, 0.6073619631901841]\n",
    "conrem_acu_le13chars = [0.5705521472392638, 0.5705521472392638, 0.588957055214724, 0.5828220858895705, 0.6073619631901841, 0.588957055214724, 0.5521472392638037, 0.558282208588957, 0.5766871165644172, 0.6012269938650306]\n",
    "conrem_acu_le14chars = [0.5950920245398773, 0.6196319018404908, 0.558282208588957, 0.6196319018404908, 0.6134969325153374, 0.5950920245398773, 0.5950920245398773, 0.588957055214724, 0.6012269938650306, 0.5950920245398773]\n",
    "conrem_acu_le15chars = [0.6073619631901841, 0.5828220858895705, 0.5828220858895705, 0.6134969325153374, 0.5828220858895705, 0.5828220858895705, 0.6012269938650306, 0.5460122699386503, 0.6134969325153374, 0.5828220858895705]\n",
    "conrem_acu_le16chars = [0.6441717791411042, 0.6134969325153374, 0.6319018404907976, 0.6073619631901841, 0.6319018404907976, 0.6196319018404908, 0.6625766871165644, 0.5950920245398773, 0.6380368098159509, 0.6012269938650306]\n",
    "conrem_acu_ac1words = [0.5521472392638037, 0.5705521472392638, 0.6073619631901841, 0.6012269938650306, 0.6073619631901841, 0.6134969325153374, 0.6012269938650306, 0.5705521472392638, 0.5705521472392638, 0.588957055214724]\n",
    "conrem_acu_ac2words =[0.7116564417177914, 0.7484662576687117, 0.7300613496932515, 0.6871165644171779, 0.7177914110429447, 0.7055214723926381, 0.7484662576687117, 0.6809815950920245, 0.7239263803680982, 0.6809815950920245]\n",
    "conrem_acu_ac3words = [0.6441717791411042, 0.6932515337423313, 0.6871165644171779, 0.6687116564417178, 0.6748466257668712, 0.6871165644171779, 0.656441717791411, 0.656441717791411, 0.6871165644171779, 0.6932515337423313]\n",
    "conrem_acu_ac3chars =[0.7116564417177914, 0.7300613496932515, 0.6993865030674846, 0.7239263803680982, 0.7055214723926381, 0.7300613496932515, 0.6380368098159509, 0.6993865030674846, 0.6871165644171779, 0.6871165644171779]\n",
    "conrem_acu_ac4chars =[0.6441717791411042, 0.6932515337423313, 0.6687116564417178, 0.656441717791411, 0.6134969325153374, 0.6993865030674846, 0.6196319018404908, 0.6503067484662577, 0.656441717791411, 0.6625766871165644]\n",
    "conrem_acu_ac5chars = [0.7239263803680982, 0.7177914110429447, 0.6748466257668712, 0.7177914110429447, 0.7177914110429447, 0.6993865030674846, 0.7361963190184049, 0.6932515337423313, 0.6932515337423313, 0.7361963190184049]\n",
    "conrem_acu_ac6chars =[0.7239263803680982, 0.7423312883435583, 0.7361963190184049, 0.6871165644171779, 0.7239263803680982, 0.7668711656441718, 0.7300613496932515, 0.7177914110429447, 0.7361963190184049, 0.7361963190184049]\n",
    "conrem_acu_ac7chars = [0.7484662576687117, 0.7116564417177914, 0.7361963190184049, 0.7300613496932515, 0.7423312883435583, 0.7239263803680982, 0.7300613496932515, 0.7116564417177914, 0.7607361963190185, 0.7607361963190185]\n",
    "conrem_acu_ac8chars = [0.754601226993865, 0.7484662576687117, 0.7239263803680982, 0.7361963190184049, 0.7239263803680982, 0.7300613496932515, 0.7300613496932515, 0.7607361963190185, 0.7177914110429447, 0.7423312883435583]\n",
    "conrem_acu_ac9chars = [0.7239263803680982, 0.7239263803680982, 0.7055214723926381, 0.7239263803680982, 0.7177914110429447, 0.7239263803680982, 0.6748466257668712, 0.6993865030674846, 0.7423312883435583, 0.7300613496932515]\n",
    "conrem_acu_ac10chars = [0.6993865030674846, 0.6687116564417178, 0.7055214723926381, 0.6748466257668712, 0.7300613496932515, 0.6993865030674846, 0.6809815950920245, 0.6993865030674846, 0.6993865030674846, 0.6871165644171779]\n",
    "conrem_acu_ac11chars = [0.6625766871165644, 0.6748466257668712, 0.6932515337423313, 0.6932515337423313, 0.6809815950920245, 0.6993865030674846, 0.7055214723926381, 0.6809815950920245, 0.6625766871165644, 0.6809815950920245]\n",
    "conrem_acu_ac12chars = [0.7177914110429447, 0.6809815950920245, 0.6871165644171779, 0.656441717791411, 0.6748466257668712, 0.6932515337423313, 0.6809815950920245, 0.656441717791411, 0.6748466257668712, 0.6687116564417178]\n",
    "conrem_acu_ac13chars = [0.6503067484662577, 0.6932515337423313, 0.6809815950920245, 0.6871165644171779, 0.6503067484662577, 0.6625766871165644, 0.6748466257668712, 0.6932515337423313, 0.6809815950920245, 0.6625766871165644]\n",
    "conrem_acu_ac14chars = [0.6687116564417178, 0.6748466257668712, 0.6319018404907976, 0.6687116564417178, 0.6932515337423313, 0.6809815950920245, 0.6871165644171779, 0.656441717791411, 0.6809815950920245, 0.6625766871165644]\n",
    "conrem_acu_ac15chars = [0.6748466257668712, 0.6932515337423313, 0.6687116564417178, 0.6625766871165644, 0.6687116564417178, 0.6503067484662577, 0.6441717791411042, 0.6748466257668712, 0.6625766871165644, 0.6687116564417178]\n",
    "conrem_acu_ac16chars = [0.6625766871165644, 0.6319018404907976, 0.6625766871165644, 0.6748466257668712, 0.656441717791411, 0.6871165644171779, 0.656441717791411, 0.6932515337423313, 0.6503067484662577, 0.6441717791411042]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plotting\n",
    "### Accuracy and standard deviation averaged over 10 trials (only raw words) depending on ngram-order (only character ngrams)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1+2\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEWCAYAAABi5jCmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nO3dd3xV9f348dc7exICWRBkiIAMBRWcleLGrZUKarWCdfysdVStbe2gtrVqa+1Qv1ar0iogiAtkuBWtAxBRkYACMgIZhJWQPd6/Pz4nyb0ZcBNyczPez8fjPnLvme97kpz3+YzzOaKqGGOMMbXCQh2AMcaYjsUSgzHGGD+WGIwxxvixxGCMMcaPJQZjjDF+LDEYY4zxY4nBmA5ARCaKyPpQx3EgInKfiPw71HGY4LLEYFpFRDaJyOmhjqO9ichjIrLPe1WISKXP58XtFEN/EXlVRHJFREUkoz32eyAicrKIvC0iu0Rkh4jMFpG0UMdlWs4SgzEtoKo3qGqCqiYA9wJzaj+r6tntFEY18CpwaTvtL1A9gYeBAcBAoAZ4PJQBmdaxxGDalIgke1ezO0Rkt/e+n8/8d0Xk9yLyPxEpEpHXRSTFZ/5VIrJZRHaKyK99SyYiMkNE/uCz7AQRyfb5/HMR2eBtd42IXOwzL1xEHhSRAhH5VkRu8q62I7z5SSLypIjkiMg2EfmDiIS34vtHiMgLIpInIntE5B0RGeYz/0IRWevFuFVEbm5mO3eKyBdNlQZUdZuqPgasDDCm33jfuUhEVovIuT7zbhCRt0TkH168G3xLgiJymM/vajGQ3Nx+VHWBqr6oqkWqWgw8ApwUSIymY7HEYNpaGPA07qqxP1CKu4r0dTkwFUgDooA7AERkBPAocAXQB0gCMluw7w3Ayd56vwOeFZE+3rxrgbOBMcDRwEUN1v0PUAUcBhwFnAn8qAX79jUfGAxkAGu9bdd6CrhKVRO9WN5vuLKI/BGYBExQ1dxWxuBrHXAi7rjcDzznm4yB8cAKoDfud+XbhjAXWOrN+wtwZQv2Ox74qvVhm1CxxGDalKruVNUXVLVEVYuAPwLfbbDY06r6taqW4k48Y7zpk4AFqvqBqlYAvwECHsxLVZ9X1e2qWqOqc4BvgGO92ZcCf1fVbFXdDdxXu56IpOOSxq2qWqyq+cBDwJRWfP8qVf2Pqu5T1TJcgjpWRGK8RaqAkSKS6B2rz3xWFxF5BHcSP11Vd7V0/83ENEdVc7zj8gywDTjGZ5F1qvpfVa3GJbEBItJTRIYCI4DfqWqFqr4FLAlknyJyDHCX9zKdjCUG06ZEJE5E/uVVBxXirjZ7NqiW8b0KLgESvPd9ga21M1S1BNjZgn1fJSKrvCqRPcAooPbK2G/bDd4PACKBHJ91/4Ur0bSIV5X0FxHZ6H3/tYDgrrjBlVQuAbZ4DbVjfVZPw5Wk/uAl1TYhItd41VK13+0w6o8LNP59gPud9AV2eAmu1uYA9jcc1wZyvap+cnDRm1CwxGDa2u3AMOA4Ve2Bq04Ad3I8kBzAtz0ilvoTKkAxEOfzOcNn2QHAE8BNQG9V7Qms9tmv37aBQ3zebwXKgRRV7em9eqjqyABibmgqrhrqFFzVzeG1IQKo6keqeh6QDrwOzPZZNw+XOGaJyLG0Ae+q/5/AdUAv77isJ/DfR4pPaQdc9eD+9jcYeAP4parObV3UJtQsMZiDESkiMT6vCCAR166wR0R6Ab9twfbmAeeLyIkiEoWrhvE9ga0CzhGRXl6j7K0+8+Jx1U47AERkKq7EUGsucIuIZIpIT3yqOFQ1B3eSflBEeohImIgMFpGGVWCBSATKcCWdeMC3sTxeRKaISA+gEijC9TCqo6qvA9OABSJyVHM78U7W0d7HaBGJbmbRBFzvoB1AmIjcgCsxBOJrXInn1yISJSKnABP3E9MA4G3gflV9OsB9mA7IEoM5GItwSaD2NR34GxALFAAfE2CdNICqfgX8BHgOd7VaBOTjruYBngE+BzbhTuRzfNZdAzwIfIS78j4C+J/P5p/w1vkC+MyLvYr6E/NVuIbwNcBuXJLqQ8s9iTsJ5wJfAh80mD8NVx2z19vnDxtuQFUXAjcAi0XkyIbzvaRQijvG4I7H3qaCUdWVwGO4xuUcYJD3/oDUPaxlMq70swv4GfDsfla5AVei+JPU39tRsJ/lTQcl9qAe01GJSAKwBxiiqt+28bbPBh5T1QFtuV1jugIrMZgORUTO9xqw43HdI7/EXREf7HZjReQcr3E4E1fF9dLBbteYrsgSg+loLgS2e68hwBRtm2Kt4NosduOqkrJw3WGNMQ1YVZIxxhg/VmIwxhjjJyLUAbRUSkqKDhw4MNRhGGNMp/Lpp58WqGpqIMt2usQwcOBAVqwIqLedMcYYj4gc8K71WlaVZIwxxo8lBmOMMX4sMRhjjPFjicEYY4wfSwzGGGP8WGIwxhjjxxKDMcYYP5YYjDHG+LHEYExH9vS57tVZdLZ4TZMsMRhjjPFjicEYY4wfSwzGGGP8WGIwxhjjxxKD6X6sgdSY/bLEYIwxxo8lBmOMMX463YN6jOlWtAZqqqAoL9SRBKa6AsIiQx2FOUiWGIwJlfIiKMyBwm1QuN29iryfhdvcvJICt+yDQ0Mba0tIOPz3Iug3DvqNhcyxEN871FGZFrDEYExbU4XS3f4n/EKfE35RjntfXth43bjekNgXevR1J9T1b0F4JJzw4/b/Hq3x4cNQWQLFBfD+X1yJByB5kH+iyDgCIqJCG2ttB4SpC0MbRwdkicGYlqiphn35TV/d+574q8r815MwSEh3J/yUITDou+59j0zo0ce9T+wDkbH+69WevMZd0z7f72CtftH9nLoQyvdBzirIXgHZy+HbpfDlXDc/PBr6HOmSReYx7mfP/iASuthNHUsMxtSqqqi/mq+92i/yrerJcZ+12n+98Ch3Uu/RFzKPrj/hJ/bxTvx9XVII72b/btEJMPA77gWuJFW4rT5RbPsUVjwNHz/q5senutJEP+/V92iI6RG6+LuxbvaXaro1Vdi1EYp3uCv6V3/qf+VfvKPxOpHx3om+LwwaX391X3vCT+zrqn/CrIPfAYlAUj/3GnmRm1ZdCXlfwbYVXsJYAV8vrl0BUg+vTxSZYyFtOISFh+wrdBeWGMzB66h1taW73VVp9qf1J57SXfXzv3qx/sq+z5j6k30Pnyv96B5WvRFM4ZHQd4x7jfuRm+b7e8teDmtfhc+ecfOiEqDvUfWJot9YSMwIXfxdlCUG0zX4XXl6J5Sd33gzBVKHweHnuJPJpzMgIhauWRLKiE1zYpPhsNPdC+pLetkrvN/vcvjwn64bL0DSIfXtFP3GQp/RjdtqTItYYjCd095tXj21VxLYvgqqSt28uBR3khg92SWCzKMhJql+3S/nhSZm0zoi0Huwe42e7KZVlkLOF/5VUGtedvPCIiB9lH8vqN6DreTXApYYTMdXUQzbP/O5YlzhGoHBNfz2GQ3HXF1fF91zgJ0EurrIWOh/nHvVKsrzSRTL4fPZsPwJNy82ub5UUXuxYJplicF0LDU1rgooe3n9lWD+mvqeQMmDXC+XzLHunzxjFEREhzZm0zEkpsPh57oXuK7FO9b5lyzfvQ9QNz8i1vV6+vo1OHSC/R35sMRgQqu4wL/ueNtnUL7XzYvu4a7yTv5pfX/3+JTQxtveOlqD/oF0pHjDwiF9hHsd80M3rbzIK30udzfjFRfArEvd39rQs2DEhTD4NIiKC23sIWaJwbSfqnLI/dKnH/sK2L3JzZMwSB8Jo75XXzfce4h1AzVtKzrRdTseNB7Wv+3uzD75p659Yu0i+PJ5iIyDIWfA8AtcsohODHXU7c4SgwkOVdizub46KHs55H7hBlkD10W031g4ZqpLBH3HQFR8aGM23Y+EuSQw5Aw4rwo2fwBr5rsusmtecXdoDz4VRlwAw852bRXdgCUG0zZqqmDDO/69RGoHgIuIdX3Pj7u+vvEvKTO08RrTUHiEa2s4dAKc82fYugyy5rtE8fVi19tp0Hddkjj8vC5drWmJwRyckl1Q8DUU58Mz3t2sKUNhyJn1vYTSRrgbmYzpLMLCYcAJ7nXWvbBtJWS94pLEglvg1dtgwEmuumn4+e6myC5EVDXUMbTI2LFjdcWKFaEOw6i6ovaiO1wDXmIfuPBh1w2wmxS3TSfXmjv2VSFvtUsQWfNhx1o3vd+xriQx/AJIHtD2scJBjzAgIp+q6thAlg1qiUFEJgJ/B8KBf6vqfQ3mPwSc4n2MA9JUtWcwYzJtoCgXFt7u6mFrh5KISoDDTgt1ZMYEl4gbMjzjCDj1btcdds18V5p4/Vfu1WeMlyQuhJTDQh1xqwQtMYhIOPAIcAaQDSwXkfmquqZ2GVW9zWf5nwBHBSse0wZU4bNn4bW7oboczrgHjv8x/PfCUEdmTGikDoPv3uleuzZC1gKXKN66x73SRrhSxIgL3QCAneTGy2CWGI4F1qvqRgAReQ64EFjTzPKXAb8NYjzmYOze5OpWN77r6lYv+KcbZsAY4/Q6FE66xb32ZkOW17Ppvfvhvfug92FekrjAlSo6cJIIZmLIBLb6fM4GjmtqQREZAAwC3m5m/nXAdQD9+/dv2yjN/tVUw7LH3dWPhMO5f3VdTO3+AtPZBfNmvKR+cPwN7lWU56pds+bD//4OH/zVPZRouNcm0W9ch/t/CmZiaCodNtfSPQWYp9rwCSjeSqqPA4+Da3xum/DMAeWvhfk/gexlrpfReQ+5P3hjTOAS090T+MZd43rxrVvkqpuWPQ4fPew6bgw/3yWJASd2iOdNBDMxZAOH+HzuB2xvZtkpQCd5qG03UF0JH/wNlj7gGpW/9wQc8f0OXfQ1plOI6wVH/cC9yva6cZrWvAIrn3GJIi7FjfU04gJ3z0SIunkHMzEsB4aIyCBgG+7kf3nDhURkGJAMfBTEWEygtq10pYS81TDqEph4PySkhjoqY7qemCQ48lL3qiiGb95w1U2rX4CV/3Hzh53jShKDT23X0IKWGFS1SkRuAl7DdVd9SlW/EpF7gBWqOt9b9DLgOe1sN1R0NZWl8M69rmibkA5TZrsH2xhjgi8q3j3udORFUFkGG952SWLdIjd8eFQCRMS4pwq2A7vBzcCmD1wpYddGOPqHcObv/R9sY4wJjaoK2LTUtUmsmuV6Pt20rFWb6jA3uJkOrqwQ3vwtrHgKkgfCVfPh0O+GOipjTK2IqPrHnBasp/n+O22823bZi+l4vn4dXr3VPQnthJvglLu7/Rj0xnRoIjTd2bPtWWLobop3wpKfw5dzIXU4XPpfN9CdMcZ4LDF0F6qut8Pin7kqpO/+HE6+3RVVjTHGhyWG7qBwuxv0bt0i6Hu0GwU1fWSoozLGdFCWGLoyVdcf+vVfu5vWzvwjHP//OsSdlcaYjssSQ1e1ayPMvxk2vQ8DT4YL/uG6uhljzAFYYuhqaqrh4/+Dt//gbqc//+/u3gQbzsIYEyBLDF1J3hqYfxNs+xSGng3n/bXd7pQ0xnQdlhi6gqoKN5Tv0r9ATA+45Ek3zpGVEozpOoI5THgDlhg6u+xPXSkhf40bAXXi/RDfO9RRGWM6MUsMnVVFCbzzR/j4UUjIgMvmwLCJoY7KGNMFWGLojL5d6ga9270Jxk6D03/nqpCMMaYNWGLoTMr2whu/gU9nuK6nVy+Egd8JdVTGmC7GEkNnsW4xvHob7MuDE2+GCb+wQe+MMUFhiaGjKy5w4xutfgHSRsKUWZB5dKijMsZ0YZYYOipV+HKeSwrlRW5Y7JNutUHvjDFBZ4mhI9qbDa/+FL55DTLHukHv0oaHOipjTDdhiaEjqamBT5+GN34LWg1n/QmOu77DD3o3dclUAJ6e+HSIIzGhZn8LwdOex9YSQ0exc4Mb9G7zBzDou26Mo16DQh2VMaYbssQQatVV8PEj8M69EB4NFzwMR/3AhrMwxoSMJYZQyl3thrPY/hkMOxfOfRB69Al1VK2i2j4PKTfGBJ8lhlCoKncD3n3wV4hNhu/PgBEXdbpSQmVNJW9tfousXVkUVxZz6txTSYtLIz0unfT4dNLj0kmLSyMjPqPufUxETKjDNsYcgCWG9rZ1GbxyExSsgyOnwMQ/QVyvUEfVInvK9jDvm3k8t/Y58kryiA6PJj0unZMyTyKvOI8tRVtYnrecooqiRusmRSe5xOElivT4dDLiMvwSSkJkAhLEJNmZGkg7U6ydjR3b5lliaC8VxfDW7+GTx6BHJlwxD4acEeqoWuTr3V8zK2sWr258lfLqco7vczy/Pv7XPL36aUSE3534O7/lSypLyC/JJ68kz72K8/zer9m5hp1lOxvtJy4iri5p1CaRhqWQ5JhkwiSsvb66Md2KJYb2sOEdWHAz7NkC466F038L0Ymhjiog1TXVvJf9HjOzZrIsdxkx4TGcP/h8rjj8Cg5LPgyAGV/NaHLduMg4BiYNZGDSwGa3X1ldSX5pPnnFeXVJJLc4t+79stxl7CjZQbVW+60XGRZZX8pokDRq36fEphARZn/ixrSU/dcEU+keeP1u+OxZ6DUYpi6GASeGOqqAFFUU8dI3LzF77Wyy92WTEZ/BbcfcxiVDLiEpOqnN9hMZHklmQiaZCZnNLlNdU82usl2NSx0lLpl8tfMr3t76NuXV5X7rhUkYKTEppMen+1VV7SzdSVR4FFsLt5IWn0Z0eHSbfZ/uqEZr6n4/e8r3EC7hbC3aSnpcOlHhdqf+wajRGnaW7iS/JJ/dZbuJi2if8dEsMQRL1quw8HYo3gHfuQ2+exdExoY6qgPatHcTs9bO4pX1r1BSVcLRaUdz2zG3cWr/U5u9+g52HW14WDipcamkxqUyKmVUk8uoKoUVheQW5/oljdpEsrlwM8tyllFUWd/ucc5L5wDQM7pn41KHz+f0uHQSohKC+h07qsqaSgpKChpVB9ZVERbnkV+aT1VNld9657zojm2vmF5+JbtGVYTx6cRHxofiq4VcZXUlO0p3NFnNWnt8d5TsoErrj23/xP7tEpslhra2Lx8W3QlrXob0I+Dy56DvUaGOar9UlQ+3f8izWc/ywbYPiAyL5OxBZ3PF8CsY0XtEqMMLiIiQFJ1EUnQSw3oNa3a5ksoSpr02jcrqSq4aeVWjBLK6YDW7ynY1Wi8+Mr7RCa62t1VtiSQ5OjmojeZtrbSqlPySfPJL8usSqu+xyC/Jp6C0AMW/K3JMeEzddz46/Wi/Y/DY549RXVPNlSOu9G9fKsnjix1fsLt8d6M4EiITDtibrWd0z051bH3b15qqIs0rzmuyfS02Irbub2xs+li/4/HY54+1W+nWEkNbUYUv5sCSn7uG5lN/DSfdAuGRoY6sWSWVJSzYsICZa2fy7d5vSYlN4cYxN/L9od8nJTYl1OEFRVxkHLERscRGxHLhYRc2uUxFdYX/P3WD6quPcz6moLSgUbtHVFgUaXFpjXtb+Vwh947tHfR2D1WlqLKI/GL/E7Pv98gvyWdv+d5G6yZGJdbFOqzXsCZLUD2iejR7kn52zbMAXDzk4ibnl1eXN4rL9xhv2L6BgtICarTGb72osKhGVYINOyb0julNeJCHj6ktmTaVRHNLcuuu9gsrChut2yOqR913GN5reOMSVHw6iZGJzR7bZ9Y8E9Tv5ssSQ1vYsxVevRXWvwmHHAcX/BNSm79qDbVt+7YxO2s2L65/kaKKIkb2Hsm937mXiQMnEtmBE1l7iQqPol9iP/ol9mt2meqaanaW7WyUNHxLHm8Vv0VFTYXfemESRkpsSqOk4XvCS4trvt2jRmvYXba7UZVDwxN/aVVpo3V7x/QmLS6NzIRMjk47utEVenpcOnGRwa3Djg6P5pAeh3BIj0OaXaaqpoqC0oKmE3NxHp/v+Jz8zflU1lT6rRcu4aTEpjRKGg2Pc3PtHr5tJX5VZg2qzxoeW0HoHeuO7SGJh/hd6fv+TmMjOn5Vci1LDAejpgZWPAlvTnclhrMfcL2OwjpeN0pVZUXeCmZmzeSdre8gCGcMOIMrhl/B6NTRnaqY3hGEh4XXlQ6O4Igml1FV9pTv8atK8D3ZbNy7kY9zPmZf5b5G6yZHJ5Men07OvhzCJIwfLv5h3cmxYX1+uLg2mPS4dIYmD+U7md9pVM2VFpvWaZJ+RFgEGfEZZMRnNLuMqrK7fHej3my1x3f9nvX8b9v/KKkqabRur5hepMels33fdsIkjKsWX9VsW0mERNQd22G9hnFyv5Prjmttck+NTe00xzZQlhhaq+Ab99zlLR/Boae4Qe+SB4Q6qkbKq8tZtHERM7Nmsm73OnpG92TaqGlMHjZ5v/94XVl73dAkIiTHJJMck7zfdo99FfvqqiIaXqFuLtxMjdYQHhbOmLQxjbrnpsel0yumV9CrUDoaEaFXTC96xfRieO/mh6TfV7Gvyd5secV5bCrcRI3WEBkW6ddW4nuMe8X06pb3y1hiaKnqKvjwH/DufRAZAxc+CmMu73DDWeSX5PPc2ueY9/U8dpfv5rCehzH9hOmce+i5NixFB5MQlUBCVAKH9jy00bzau3OfOuup9g6rS6g9toN7Dm40r/bYPnnWk+0dVodniaElcr5wg97lfA7Dz4dzHoTE9FBH5eeLHV/wbNazvLHpDaq1mgmHTOAHw3/AuIxxVl1kjAmIJYZAVJbB0gfgg79BXG+49L8woukeLaFQWVPJG5veYGbWTL4o+IKEyAQuG34Zlx1+GYckNt/IZ4wxTbHEcCBbPnaD3u38BkZfDmf9scMMererbBfPr3ueuevmkl+az8AeA/nFsb/gwsMu7LY3DRljDl73SQxPn+t+Tl0Y2PLl++Cte2DZ45B0CPzgBTjs9ODF1wLrdq3j2axnWbRxERU1FZzU9ySmnzidkzJP6pYNZabjsJFKg6c9j21QE4OITAT+DoQD/1bV+5pY5lJgOqDA56p6eTBjCsj6t2DBrbB3Kxx7HZz2G4gO7ZAI1TXVvLP1HWZmzWRF3gpiI2K5eMjFXH745U02WhpjTGsFLTGISDjwCHAGkA0sF5H5qrrGZ5khwC+Ak1R1t4ikBSuegJTsgtd/BatmQu8hMG0J9D8+pCHtLd9bN5jd9uLt9I3vy+3H3M7FQy5u08HsjDGmVjBLDMcC61V1I4CIPAdcCKzxWeZa4BFV3Q2gqvlBjGf/1rwCC++Akp1w8u0w/meuO2qIbNy7kVlZs5i/YT6lVaWMTR/Lz8b9jAmHTOh2fdaNMe0rmIkhE9jq8zkbOK7BMkMBROR/uOqm6aq6JIgxNVaUC4vugKwFkHGka0voc2S7hlCrRmv4YNsHzMyayYfbPyQqLIpzDj2HK4ZfweG9Dg9JTMY0p7KykuzsbMrKykIdSqtMS5sGQFZWVogjaVsxMTH069ePyMjW340dzMTQVKf5hk+MjwCGABOAfsD7IjJKVff4bUjkOuA6gP7922jYWVVYNQte+4Xrjnrab+HEn4Rk0LviymJeWf8Ks9fOZlPhJtJi0/jJUT9h0tBJ9IrpGD2gTGh05Mbc7OxsEhMTGThwoN0j00GoKjt37iQ7O5tBgwa1ejvBTAzZgG8n+n7A9iaW+VhVK4FvRWQdLlEs911IVR8HHgcYO3Zsw+TScrs3w4JbYOM70P8EN+hdypCD3mxLbS3ayuy1s3npm5fYV7mPI1OO5P6T7+eMAWd0ubFXTNdTVlZmSaGDERF69+7Njh07Dmo7wUwMy4EhIjII2AZMARr2OHoZuAyYISIpuKqljUGLSBU++Re8+Ts3hMU5f4Gx17TroHeqyrLcZTyb9SzvbX2PcAnnzIFncsXwKzgyNTRVWMa0liWFjqctficHTAwichMws7aBOFCqWuWt+xqu/eApVf1KRO4BVqjqfG/emSKyBqgG7lTVxk+vaAuVJe4mtS3/c/cjnPcQ9GyfpyEBlFWVsXDjQmaunck3u78hOTqZa4+8lsnDJpMWF9rOWMa0l8n/+giAOdefEOJIzP4EUmLIwHU1XQk8BbymqgFV56jqImBRg2m/8XmvwE+9V1DNLt3C+ZXFJFz8LzhycrsNepdbnMucdXOY9/U89pTvYVjyMO458R7OOfQce9awMQchNzeXW2+9leXLlxMdHc3AgQP529/+xtChQ9tk+xMmTCAnJ4eYmBgSEhJ46qmnGDZsmN/0qKgonnjiCcaMGdMm+wSYPn06TzzxBKmpqVRVVXHvvfdywQUX1M0fPXo0I0aMYPbs2W22z4YOWIeiqr/C1fs/CVwNfCMi94pI4+EKO7A3evTkl5kDYPSUoCcFVWVV/irueO8OJr4wkadWP8XY9LE8ddZTPH/+81w85GJLCsYcBFXl4osvZsKECWzYsIE1a9Zw7733kpeX16b7mTlzJp9//jk//OEPufPOOxtNv/HGG/2mt5XbbruNVatW8fzzzzNt2jRqatwT7bKysqipqWHp0qUUFxe3+X5rBVS57l3Z53qvKiAZmCciDwQtsjamIhSFB3cEkMrqShZsWMBlCy/jysVX8uH2D7lyxJUs+t4iHjrlIRvh1Jg28s477xAZGckNN9xQN23MmDGcfPLJfsvdddddPProo3Wfp0+fzoMPPkhOTg7jx49nzJgxjBo1ivfff3+/+xs/fjzr169vNP2EE05g27ZtjaYvXryYSy+9tO7zu+++y/nnn091dTVXX301o0aN4ogjjuChhx7a736HDx9OREQEBQUFAMyaNYsrr7ySM888k/nz5+933YMRSBvDzcAPgQLg37h2gEoRCQO+AX4WtOg6iYLSAjeY3ddzKSgtYFDSIH513K84f/D5QX9UojEdwe8WfMWa7Y2fc9zQmhy3TG1bw/6M6NuD354/ssl5q1ev5phjjjngNqZMmcKtt97KjTfeCMDcuXNZsmQJs2bN4qyzzuLuu++murqakpLGT3rztWDBAo44ovGT+pYsWcJFF13UaPoZZ5zB9ddfT3FxMfHx8cyZM4fJkyezatUqtm3bxurVqwHYs2dPo3V9ffLJJ4SFhZGamgrAnDlzeOONN1i3bh0PP/wwl1122QGPQWsEcgmdAnxPVTf7TlTVGhE5LyhRdRJrdq5hZtZMFn+7mMqaSk7OPJkrhl/BCX1PsMHsjOkAjjrqKPLz85Dx/voAAB7FSURBVNm+fTs7duwgOTmZ/v37M27cOKZNm0ZlZSUXXXRRs20EV1xxBbGxsQwcOJB//vOfftOLi4uprq5m5cqVjdaLiIhg4sSJLFiwgEmTJrFw4UIeeOABqqqq2LhxIz/5yU8499xzOfPMM5vc70MPPcSzzz5LYmIic+bMQURYvnw5qampDBgwgH79+jFt2jR2795NcnJy2xws3/gDWGYRsKv2g4gkAiNU9RNV7Vq3DAagqqaKt7e8zcysmazMX0lcRByThk7i8sMvZ2DSwFCHZ0xINHdl31Bb9UoaOXIk8+bNC2jZSZMmMW/ePHJzc5kyZQrgqoaWLl3KwoULufLKK7nzzju56qqrGq07c+ZMxo4d2+T00aNH8/Of/5wf//jHvPjii42WmTx5Mo888gi9evVi3LhxJCYmAvD555/z2muv8cgjjzB37lyeeqrx0/luu+027rjjDr9ps2fPZu3atQwcOBCAwsJCXnjhBX70ox8FdBxaIpDL2v8DfJ9WXuxN61b2lu/lyS+f5OwXz+b2924nrySPO8feyZvff5NfHvdLSwrGtKNTTz2V8vJynnjiibppy5cv57333mu07JQpU3juueeYN28ekyZNAmDz5s2kpaVx7bXXcs011zR51X8gkZGR/OEPf+Djjz9ucliNCRMmsHLlSp544gkmT54MQEFBATU1NVxyySX8/ve/D3i/NTU1PP/883zxxRds2rSJTZs28corrwStZ1IgJQbx7Z7qVSF1m+c4rN+9nplrZ/Lqhlcpqy7juIzj+OWxv2R8v/E2mJ0xISIivPTSS9x6663cd999xMTE1HVXbWjkyJEUFRWRmZlJnz59ANcY/Oc//5nIyEgSEhL473//26o4YmNjuf322/nLX/7Ck0/6Pzs6PDyc8847jxkzZvCf//wHgG3btjF16tS6XkZ/+tOfAtrP0qVLyczMJDMzs27a+PHjWbNmDTk5OXXfq63IgW5JEJEXgXepLyXcCJyiqo1bXNrB2LFjdcWKFS1eb+oMVxx8+uoDr1ujNbyf/T7PZj3LxzkfEx0ezXmHnsflwy9naHLb9JE2prPLyspi+PDhLVrHbnBrH039bkTkU1VtXC/WhECu/G8A/gH8CjcI3lt4A9p1Nfsq9vHy+peZtXYWW4u2khaXxi1H38IlQy4hOabtG3iM6W4sIXQOB0wM3jMSprRDLMGV0birWa0thVuYtXYWL69/meLKYsakjuHmo2/mtP6nERlmg9kZY7qXQO5jiAGuAUYCdU+uUdVpQYwr6FSVj3I+YlbWLJZmLyU8LJyJAyfyg+E/YGRKYD0sjDGmKwqkKukZYC1wFnAPcAXQabupllaVsmDDAmZlzWLD3g30iunF9aOv59Khl5Ialxrq8IwxJuQCSQyHqer3ReRCVf2PiMzCjYraqZRXl7OjZAenP386hRWFDO81nD9+549MHDiRqPCoUIdnjDEdRiCJodL7uUdERuHGSxoYtIiCZFfZLnJLcjljwBn8YPgPOCrtKBu3yJj29vS57ufUhaGNw+xXIDe4PS4iybheSfOBNcD9QY0qCFJjUzki5Qj+OuGvHJ1+tCUFYzq52juZBw8ezIgRIzjnnHP4+uuvA1p3woQJDBs2jNGjR3PSSSexbt26RtPHjRvHqlWr2jTm6dOnk5mZWTd4X8OB8EaPHh208Y9aYr+JwRsor1BVd6vqUlU9VFXTVPVf7RRfm4kIi7Chro3pIgIddnvGjBlMnz69yW101yG1A7HfxKCqNcBN7RSLMcYEJNBhtwPR3YbUDkQgbQxviMgdwBzcOEkAqOqu5lcxxnQri38OuV8eeLncL9zP2raG/ck4As6+r8lZgQ67HYjuNqR2IAJJDLX3K/zYZ5oCh7Z9OMYYc3B27tzJaaedBsCuXbuoqKjg5ZdfBuCZZ56pSwLddUjtQARy5/Og9gjEGNOJNXNl30gb9Ura37DbvXv3rms0njFjBps2bWqynaG7DqkdiAP2ShKRq5p6tUdwxhjTlJYMu90aXXlI7UAE0l11nM/rZGA6cEEQYzLGmP2qHXb7jTfeYPDgwYwcOZLp06fTt2/fNtuH75DaDdUOqb148WLOO889yHLbtm1MmDCBMWPGcPXVV7fZkNqhcMBhtxutIJIEPKOqIUkOrR52e8lUAJ6e+HRbh2RMt9SaYbftBrf20R7DbjdUAgxpxXrGmO7OEkKnEMjoqgtwvZDAVT2NAOYGMyhjjDGhE0iJwbeCrQrYrKrZQYrHGGNMiAWSGLYAOapaBiAisSIyUFU3BTUyY4wxIRFIYngeONHnc7U3bVxQIgoSa3Q2xpjABNJdNUJVK2o/eO/tAQbGmBabumRqXQ9B03EFkhh2iEhd11QRuRAoCF5IxhhzYAcz7HYggjE0d5cYdttzA/BLEdkiIluAu4DrgxuWMcY0L9Bhtw9Wa4fm3rRpExMmTGhym51+2G0AVd2gqsfjuqmOVNUTVbXxGLXGGNNOAh12+6677uLRRx+t+zx9+nQefPBBcnJyGD9+fN2V+/vvv7/f/bV0aO5AdOpht0XkXuABVd3jfU4GblfVXwU7OGNM53D/svtZu2vtAZerXSaQdobDex3OXcfe1eS8QIfdnjJlCrfeeis33ngjAHPnzmXJkiXMmjWLs846i7vvvpvq6mpKSkr2u52WDs0diM4+7PbZqvrL2g+qultEzsE96tMYYzqso446ivz8fLZv386OHTtITk6mf//+jBs3jmnTplFZWclFF13EmDFjmly/pUNzX3zxxXz77bdUVFSwZcuWuu3ecsstTJ3qkmGXGHYbCBeRaFUtB3cfA2DPyDTG1Gnuyr6hthqzbH/Dbjc0adIk5s2bV9dYDa5qaOnSpSxcuJArr7ySO++8k6uuajxodEuH5n7ppZcA18Zw9dVX8+677zZat0sMuw08C7wlIteIyDXAG8B/ghuWMcY0ryXDbk+ZMoXnnnuOefPmMWnSJAA2b95MWloa1157Lddcc03AQ2T7OtDQ3IHqlMNuq+oDwB+A4bgG6CXAgCDHZYwxzWrJsNsjR46kqKiIzMxM+vTpA7jnNI8ZM4ajjjqKF154gVtuuaVVcexvaO5Addpht0VkDHA5cCnwLfCCqj4c5Nia1Npht40xbas1w27b8PftI2jDbovIUGAKcBmwE5iDSySnBBqciEwE/g6EA/9W1fsazL8a+DNQ29/rYVX9d6DbN8Z0LpYQOof9NT6vBd4Hzq+9b0FEbgt0wyISDjwCnAFkA8tFZL6qrmmw6BxVvallYRtjjAmW/bUxXALkAu+IyBMichogLdj2scB6Vd3oja/0HHBh60M1xnQ0LX0CpAm+tvidNJsYVPUlVZ0MHA68C9wGpIvI/4nImQFsOxPY6vM525vW0CUi8oWIzBORQwIP3RgTSjExMezcudOSQweiquzcuZOYmJiD2s4B72NQ1WJgJjBTRHoB3wd+Drx+gFWbKl00/AtaAMxW1XIRuQHXDfbURhsSuQ64DqB///4HCtkY0w769etHdnY2O3bsCHUoxkdMTAz9+vU7qG0E1CupVRsWOQGYrqpneZ9/AaCqf2pm+XBgl6om7W+71ivJGGNariW9kgK5wa21lgNDRGSQiEThejj5jQwlIn18Pl4AtP4uEWOMMW0ikCExWkVVq0TkJuA1XHfVp1T1KxG5B1ihqvOBm71nPVQBu4CrgxWPMcaYwAStKilYrCrJGGNarqNUJRljjOmELDEYY4zxY4nBGGOMH0sMxhhj/FhiMMYY48cSgzHGGD+WGIwxxvixxGCMMcaPJQZjjDF+LDEYY4zxY4nBGGOMH0sMxhhj/FhiMMYY48cSgzHGGD+WGIwxxvixxGCMMcaPJQZjjDF+LDEYY4zxY4nBGGOMH0sMxhhj/FhiMMYY48cSgzHGGD+WGIwxxvixxGCMMcaPJQZjjDF+LDEYY4zxY4nBGGOMH0sMxhhj/HSbxDD5Xx8x+V8fhToMY4zp8LpNYjDGGBMYSwzGGGP8WGIwxhjjxxKDOWidrf2ms8VrTHuzxGCMMcaPJQZjjDF+LDEYY4zxExHqANpLwb5y8grLueqpZSRGR5AQHUFCTATx0RHuc0z9tNrP8VERJHrLRIZbDjXGdA/dJjGICGECe0sr2ba7hH3lVewrq6K4ojqg9WMiw0iIjiTRSyDx0eF+n2sTS/38phJOJDGRYYhIkL+t6SpqG8nnXH9CiCMxodaefwvdJjH0jo+id3xUo4NaXaMUV1RR7CWKIu/nPt+f3quobnol+8qr2LanlH3llW69siqqavSAcYSHCfFR4STGRPollNqSSny0f4LxTTh186IjiY8OJ8JKMcaYIAhqYhCRicDfgXDg36p6XzPLTQKeB8ap6opgxtRQeJjQIyaSHjGRkNT67agq5VU1jROK974+4VRSXF7tJRmXYPaUVLB1d4krwZQHXoqJjQxvlFB8q8L8Ek+0f8nFd7qVYowxvoKWGEQkHHgEOAPIBpaLyHxVXdNguUTgZuCTYMXSHkSEmMhwYiLDSUmIPqht1ZZi6pKK97Nxqaaybn6xl4i27irxS0qBlmISmimp+CaU+vm1icVVp5VXVRMRZqUXY4KpsrqGsHa6gAtmieFYYL2qbgQQkeeAC4E1DZb7PfAAcEcQY+lU/EoxB6GpUkxddZhXBbavvLq+OsxnuV3FFWzZVVL3uSSAUsyo375GRlIMfZJiSO/hfmYkxZDRI8abHktyXKSVTowJUH5hGa99lcvi1bms3LKHQSnx7bLfYCaGTGCrz+ds4DjfBUTkKOAQVX1VRJpNDCJyHXAdQP/+/YMQatfUlqWYquoaiiuqm6wq+8vr66iqruG04enk7i0jt7CMb/IKyC8qo2GBJSoizCdRxPi9d8kkltTEaMLDLHl0Rp2psbyjxrptTymLv8xhyepcPt2yG1UYnBpP36QYEqPbp1k4mHtp6j+77jQhImHAQ8DVB9qQqj4OPA4wduzYA9eNmDYXER5GUmwYSbGNSzH//WgTANMvGOk3vaq6hoJ9FeTsLa1LGLl7y8jZ635+tmUPuXvLqKiu8VsvPExIS4xuttTRJymGtB7RREeEB+vrGtOuNhUUs3h1LktW5/B59l4ADs9I5NbThnLOERkMSU9s12FcgpkYsoFDfD73A7b7fE4ERgHvelULGcB8EbmgvRugTXBEhIe5k3pSTLPLqCq7iisaJY3az1/nFbH06x1NNsj3jo9qkDTqSx21+01opyssY1rqm7wiFn2Zy+LVOazNLQJgdL8k7pp4OGePymBgO1UbNSWY/zXLgSEiMgjYBkwBLq+dqap7gZTazyLyLnCHJYXuRUTonRBN74RoRvZtvltYUVllXcKoTR45e8vIKyxj+94yVm7Zze6SykbrJUZH1CWJDK8EkldYRpgIL3yaHcyv1iZ2FJUTJrB80y4yelhJqTNTVb7aXsiS1S4ZbNhRjAgc0z+ZX507nImjMuiXHBfqMIEgJgZVrRKRm4DXcN1Vn1LVr0TkHmCFqs4P1r6b0tHqEU3LJMZEkhgTyZD0xGaXKausJq+wcakjZ28puYXlfJ23g/yictSrjLz9+c/bKfqD9/3H6qsRUhIalpRiG1W7xVtJqUOoqVFWZe+pSwZbd5USJnD8ob25+sSBnDUyg7QezZeofbXnOSyofz2qughY1GDab5pZdkIwYzFdX0xkOAN6xzOgd/NF8KrqGiY99hGqyj8vO7odo2udn8xeSY0qd551eH1VW6Frs8neXcqKzbvZ01RJKSbCSxSxZPSIJsNrm/EtOSXFWg+xYKiuUVZs2uW1GeSSW1hGZLhw0mEp3HTKYZwxIoNe8VGhDnO/7LLCdCsR4WFER7h7Lvr37hjF9v2JiXTVRuOHpja7TFlldYOkUU7u3tK6qra1OYXs2FdfUqrfdphfqaNhe01GjxhSEqIJsx5iB1RZXcPHG3eyeHUur3+VS8G+CqIjwhg/NJWfjRrGacPTm+y40VFZYjCmk4uJDGdgSvx+Gysrq2vILyp3VWxe9Zpvtduyb3eRX1RGZbV/9ogIE9J7NOwZ5n+fSlpiDFER3e8Gx/Kqaj74poDFq3N5MyuPPSWVxEWFc8rhaZw9KoNThqV12iq9zhm1MaZFIsPDyOwZS2bP2GaXqalRdhZX+LTPlPq112TlFPL22nxKK/17iIlA7/ho+iTFkL27hMjwMB55Z32jrsZxUZ3/dFNaUc17X+ezeHUub2flU1ReRWJMBKcPT2fiqAy+OzS1rpTXmXX+35QJOWvY7xrCwoTUxGhSE6M5opmBw1SVwtIqr3dY4/tT1ufvo7Csij+/tq7Rukmxkf5VVQ0a0DOSYugRE9Hh2j2Kyip5e20+S1bn8u66HZRWVpMcF8k5R/Rh4hEZnDQ4pcuVmCwxGGMCJiIkxUWSFBfJsIzGPcRqb8J6euq4Jm9qrP381fZCCvaVN1o/NjK8iaRR24juPveOjwp6u8fekkreyMpj8Zc5vP9NARXVNaQmRjPpmH6cPSqDYwf16tKjG1tiMMa0ubioCA5NTeDQ1IRml6moqiG/yP+eFN8E8sm3u8grLGs0EGRkuNfu0cyNjX2SYkhNjG7xw7UK9pXz+ld5LF6dw0cbdlJVo/RNiuEHxw/g7CMyOKZ/crdpiLfEYIwJiaiIMPolx+33pq7qGmXnvnK/Gxt9709ZvW0vb2blUVbpP6yKCKQmRDcofcT6NZzX1ChVNcqM/33L4tW5LN+0ixqFgb3j+NHJh3L2qAyO7JfU4aq22oMlBmNMhxUeJqT1iCGtRwxH9mt6GVVlb2ml113XpwSyt4ycwjK+LSjmww07KSqranL9z7buYUhaAjedchhnH9GHwzMSu2Uy8GWJwZgOzBr2D0xE6BkXRc+4KIb36dHscsXlVXWJI3dvGQ+9+TUCPD31WA5La77KqzuyxGCM6RbioyMYnJrAYK/dY+4K91QASwqNdd1mdWOMMa1iicEYY4wfSwzGGGP8WBuD6XasQTd47Nh2DVZiMMYY48cSgzHGGD+WGIwxxvixxGCMMcaPJQZjjDF+LDEYY4zxY4nBGGOMH0sMxhhj/NgNbsaYbsluxmuelRiMMcb4scRgjDHGjyUGY4wxfiwxGGOM8WOJwRhjjB9LDMYYY/xYYjDGGOPHEoMxxhg/lhiMMcb4EVUNdQwtIiI7gM2tXD0FKGjDcEy9znZsO1O8nSlW6FzxdqZY4eDiHaCqqYEs2OkSw8EQkRWqOjbUcXRFne3YdqZ4O1Os0Lni7UyxQvvFa1VJxhhj/FhiMMYY46e7JYbHQx1AF9bZjm1nirczxQqdK97OFCu0U7zdqo3BGGPMgXW3EoMxxpgDsMRgjDHGT6dKDCLylIjki8jqdt7vBBF5tT33GSoisklEvhSRVSKyogXrzRCRScGMrYl99hSReSKyVkSyRCSgR3KFKNZh3jGtfRWKyK0BrhuKeG8Tka9EZLWIzBaRmADXe1dE2r37p4jc4sX6VaDH1VuvXY5tU+cuEeklIm+IyDfez+QWbO9qEXk4ONF2ssQAzAAmBnsnIhIe7H10cKeo6pj26t8tTmv+Fv8OLFHVw4HRQFbbRtZYa2NV1XXeMR0DHAOUAC+1eYANtCZeEckEbgbGquooIByYEoz4Guy3Vf93IjIKuBY4Fvd3cJ6IDGnL2JrZb0uO7Qwan7t+DrylqkOAt7zP7eJAx7pTJQZVXQrsam6+iPxMRG723j8kIm97708TkWe995d5V8SrReR+n3X3icg9IvIJcIKITPSuRD8AvhfUL9bJiMhVIvKFiHwuIs/4zBovIh+KyMbaqzARSRCRt0RkpXfcL/SmD/Su8h8FVgKHtDCGHsB44EkAVa1Q1T0dMdYmnAZsUNVGd/B3oHgjgFgRiQDigO1NxPozb7+fi8h9PrO+LyLLRORrETnZJ6b3vVhXisiJ3vQJIvKOiMwCvmxFnADDgY9VtURVq4D3gIubiDdkx7aZc9eFwH+89/8BLmpqXREZ58X3uXdcE71ZfUVkibgSxwM+y/+fiKwQV3r6nc/0TSLyG++c9v0DBdypXsBAYHUz844Hnvfevw8sAyKB3wLXA32BLUAq7g//beAib3kFLvXexwBbgSGAAHOBV0P93dvp+H6L+4P/FLiuifkjgXVAive5l/dzBvA87mJjBLDemx4B9PDepwDrvWM6EKgBjm9lnGO83+8M4DPg30B8R4y1idifAm7qqMfW294twD5gBzCziflnAx8CcQ1ifRd40Ht/DvCm9z4OiPHeDwFWeO8nAMXAoIOIdTjwNdDb289HwD872rGlwbkL2NNg/u4m1okCNgLjvM89vNiu9qYn4c5Xm4FDGny3cO/3caT3eRPws0Bi7VQlhgB8ChzjZdRy3B/IWOBkXKIYB7yrqjvUXVnMxF11AlQDL3jvDwe+VdVv1B3RZ9vxO4TaSap6NO4f/8ciMr7B/FOBeapaAKCqvldBL6tqjaquAdK9aQLcKyJfAG8CmT7zNqvqx62MMwI4Gvg/VT0Kd3JpWBTvKLHWEZEo4ALcyaihDhGvuLruC4FBuIupeBH5QYPFTgeeVtWSJmJ90fv5Ke5kCO4C7QkR+RL33Uf4LL9MVb9tTazevrOA+4E3gCXA50BVg8U6xLFthWFAjqou9+Iu9M5d4Kqh9qpqGbAGGOBNv1REVuIumEbif6znBLLTTp0YROQQqW/Mu0FVK3FZcSruauZ94BRgMK7+WfazuTJVrfb53C1v8FDV7d7PfFwd+Hd9jzHuGDZ3bMp93tce6ytwJbRj1NWv5+GucMCdzFsrG8hW1U+8z/OAEztorL7OBlaqal7Dv98OFO/puAujHd7/1IvAFT6xXhBgrNW4BA5wmxffaNzFWpTP8gd9bFX1SVU9WlXH46ps9nTQY+srT0T6AHg/8733r3lx/7sFcVcDESIyCLgDOE1VjwQW+sQdcOydOjGo6lb1GvRU9TFv8lLcgVmKSww3AKu8K/9PcCe6FHGNL5fh6iMbWgsMEpHB3ufLgvpFOggRia+tvxSReOBMYHmDY/wW7oqkt7dcrwNsNgnIV9VKETmF+quag6KqucBWERnmTToNd8LtcLE2cBkw2/sODf9+O0q8W4DjRSRORAR3bBf5xDofeB2YJiJxLYg1R1VrgCtx1RxtRkTSvJ/9cW2C/+igx9bXfOCH3vsfAq8AqOpZXtw/wp2L+orIOC/uRK/dpzk9cCf/vSKSjrsQabH97aDDEZHZuDrJFBHJBn6rqk82WOx94G7gI1UtFpEybxqqmiMivwDewWXiRar6SsP9qGqZiFwHLBSRAuADYFSwvlcHkg685M4FRACzVHWJ7wKq+pWI/BF4T0SqccXVq/ezzZnAAnFdX1fh/tDbyk+AmV71zEZcSbGjxop3Ej0D197VSEeJV1U/EZF5uLamKi+Oxxsss0RExgArRKQCWAT8cj+bfRR4QUS+j/v/a6ur7loveCf9SuDHqrq7QbwhPbZNnbuA+4C5InINLhk3ahBW1QoRmQz8U0RigVJcia5Jqvq5iHwGfIX7n/hfq+L1GiWMMcYYoJNXJRljjGl7lhiMMcb4scRgjDHGjyUGY4wxfiwxGGOM8WOJwXQ5IqIi8qDP5ztEZHoIQ2oVEZkuIneEOg7T/VhiMF1ROfA9EUlpzcoSgtF122KfB7jxyZiAWWIwXVEV7oas2xrOEJHBIvKxiCwXN5ruPm96o1E+ReRlEflU3CiV1/lsY5+I3O/Ne1NEjhX3HIKN3nARDfcpIvJncSP6fundsNTcPu8WkXUi8iZunBzfuJd4+3xfRA73ps8Qkb+KyDu48YKMOWh2hWG6qkeAL8RnOGLP34G/q+psbwwdX8cCo3wGdJumqru8O06Xi8gLqroTiMcNxniXiLwE/AF3R/MI3PDJ8xts93u40WBH40bqXC4iSxvuU0SOwT334Cjc/2btKLfgEt0NqvqNiByHu5P4VG/eUOD0BmN9GdNqlhhMl6SqhSLyX9wDZ0p9Zp1A/bj3s4C/+MxrOMrnzSJSO67/IbjhoncCFbhRPMFd6Zd7Y+p8Sf1oor6+A8z2Ttx5IvIebqTfwgb7PBl4qXbEUhGZ7/1MAE4EnveGKwGI9tn+85YUTFuyxGC6sr/hrrqfDnD5uvF7RGQCbkyaE1S1RETepX6UykqtH0umBm+US1Wtaaaef3+j+jYcM6ipMWrCcGP3jwlwG8YcFGtjMF2WN+b+XOAan8kfA5d47/f3uMok3INTSrz6/OMPIpSlwGQRCReRVNwzQJY1s9zFIhIrbpTb873vUQh86w1AV9tmMfog4jFmvywxmK7uQVy9fq1bgZ+KyDKgD7C3mfWW4Ma3/wL4PS6htNZLwBe4B8i8jXuKVm7DhVR1Je5BKqtwD41632f2FcA1IvI5buTMCw8iHmP2y0ZXNd2KN/R1qaqqiEwBLlNVO8ka48PaGEx3cwzwsLhW3D3AtBDHY0yHYyUGY4wxfqyNwRhjjB9LDMYYY/xYYjDGGOPHEoMxxhg/lhiMMcb4+f9kxgZQwckRWAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "print(\"Taaloefening 1\")\n",
    "# example data\n",
    "y_c_vs_pr_vs_pa = [.35,.31, .30,.27,.27,.29,.32]\n",
    "yerr_c_vs_pr_vs_pa = [.03,.04,.03,.02,.04,.03,.03]\n",
    "x_values = [\"1-word\", \"5-char\", \"6-char\", \"7-char\", \"8-char\", \"9-char\", \"10-char\"]\n",
    "\n",
    "y_cpr_vs_pa =  [.55,.64,.65,.65,.66,.65,.64]\n",
    "yerr_cpr_vs_pa = [.02,.04,.02,.02,.02,.02,.02]\n",
    "\n",
    "y_c_vs_prpa = [.52,.61,.66,.68,.67,.68,.65]\n",
    "yerr_c_vs_prpa = [.04,.03,.03,.03,.03,.02,.03]\n",
    "\n",
    "x = [3,5,6,7,8,9,10]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(\"Ngram order\")\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.errorbar(x, y_c_vs_pr_vs_pa, yerr=yerr_c_vs_pr_vs_pa, label = \"C vs PR vs PA\")\n",
    "ax.errorbar(x, y_cpr_vs_pa, yerr=yerr_cpr_vs_pa, label = \"C+PR vs PA\")\n",
    "ax.errorbar(x, y_c_vs_prpa, yerr=yerr_c_vs_prpa, label = \"C vs PR+PA\")\n",
    "plt.xticks(x, x_values)\n",
    "plt.legend()\n",
    "plt.title(\"Language Task 1\")\n",
    "plt.savefig(\"ngram_tracing_taaloefening1\")\n",
    "plt.show()\n",
    "print(\"Taaloefening 2\")\n",
    "\n",
    "y_c_vs_pr_vs_pa = [.42,.45, .51,.51,.51,.49,.49]\n",
    "yerr_c_vs_pr_vs_pa = [.05,.04,.03,.05,.04,.04,.05]\n",
    "\n",
    "y_cpr_vs_pa =  [.70,.75,.81,.82,.79,.77,.73]\n",
    "yerr_cpr_vs_pa = [.03,.04,.04,.03,.02,.02,.03]\n",
    "\n",
    "y_c_vs_prpa = [.56,.62,.62,.59,.58,.61,.60]\n",
    "yerr_c_vs_prpa = [.04,.02,.03,.03,.03,.02,.04]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(\"Ngram order\")\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.errorbar(x, y_c_vs_pr_vs_pa, yerr=yerr_c_vs_pr_vs_pa, label = \"C vs PR vs PA\")\n",
    "ax.errorbar(x, y_cpr_vs_pa, yerr=yerr_cpr_vs_pa, label = \"C+PR vs PA\")\n",
    "ax.errorbar(x, y_c_vs_prpa, yerr=yerr_c_vs_prpa, label = \"C vs PR+PA\")\n",
    "plt.title(\"Language Task 2\")\n",
    "plt.xticks(x, x_values)\n",
    "plt.legend()\n",
    "plt.savefig(\"ngram_tracing_taaloefening2\")\n",
    "plt.show()\n",
    "print(\"Taaloefening 1+2\")\n",
    "\n",
    "y_c_vs_pr_vs_pa = [.41,.40, .39,.39,.38,.37,.40]\n",
    "yerr_c_vs_pr_vs_pa = [.02,.02,.03,.02,.02,.03,.01]\n",
    "\n",
    "y_cpr_vs_pa =  [.59,.71,.73,.74,.74,.72,.69]\n",
    "yerr_cpr_vs_pa = [.02,.02,.02,.02,.02,.02,.02]\n",
    "\n",
    "y_c_vs_prpa = [.56,.64,.63,.62,.63,.62,.63]\n",
    "yerr_c_vs_prpa = [.03,.01,.01,.02,.02,.02,.02]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.set_xlabel(\"Ngram order\")\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.errorbar(x, y_c_vs_pr_vs_pa, yerr=yerr_c_vs_pr_vs_pa, label = \"C vs PR vs PA\")\n",
    "ax.errorbar(x, y_cpr_vs_pa, yerr=yerr_cpr_vs_pa, label = \"C+PR vs PA\")\n",
    "ax.errorbar(x, y_c_vs_prpa, yerr=yerr_c_vs_prpa, label = \"C vs PR+PA\")\n",
    "plt.xticks(x, x_values)\n",
    "plt.legend()\n",
    "plt.title(\"Language Task 1 and 2\")\n",
    "plt.savefig(\"ngram_tracing_taaloefening12\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ngrams = df_ngrams.loc[~df_ngrams.index.duplicated(keep='first')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "steps = [0.0,0.01,0.02,0.03,0.04,0.05,0.06,0.07,0.08,0.09,0.1,0.11,0.12,0.13,0.14,0.15,0.16,0.17,0.18,0.19,0.2]\n",
    "controls = []\n",
    "controls_correct = 0\n",
    "acute = []\n",
    "acute_correct = 0\n",
    "remission = []\n",
    "remission_correct = 0\n",
    "for participant,row in df_ngrams.iterrows():\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 0:\n",
    "        controls.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "        if df_ngrams.loc[participant, \"Predicted Label\"] == 0:\n",
    "            controls_correct += 1\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 1:\n",
    "        acute.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "        if df_ngrams.loc[participant, \"Predicted Label\"] == 1:\n",
    "            acute_correct += 1\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 2:\n",
    "        remission.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "        if df_ngrams.loc[participant, \"Predicted Label\"] == 2:\n",
    "            remission_correct += 1\n",
    "\n",
    "y = []\n",
    "y1 = []\n",
    "y2 = []\n",
    "\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in controls:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y.append(0)\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in acute:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y1.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y1.append(0)\n",
    "y2 = []\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in remission:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y2.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y2.append(0)\n",
    "print(\"Control vs Acute vs Remission\")\n",
    "plt.plot(steps,y, label = \"Control\")\n",
    "plt.plot(steps,y1, label = \"Acute\")\n",
    "plt.plot(steps,y2, label = \"Remission\")\n",
    "plt.xlabel(\"Overlap Coefficient Difference between 1st and 2nd highest Coefficent\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "steps = [0.0,0.01,0.02,0.03,0.04,0.05,0.06,0.07,0.08,0.09,0.1,0.11,0.12,0.13,0.14,0.15,0.16,0.17,0.18,0.19,0.2]\n",
    "controls = []\n",
    "patients = []\n",
    "for participant,row in df_ngrams.iterrows():\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 0:\n",
    "        controls.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 1:\n",
    "        patients.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 2:\n",
    "        patients.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "y = []\n",
    "y1 = []\n",
    "\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in controls:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y.append(0)\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in patients:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "            elif df_ngrams.loc[tupl[1], \"Predicted Label\"] == 1:\n",
    "                lstcorre.append(tupl)\n",
    "            elif df_ngrams.loc[tupl[1], \"Predicted Label\"] == 2:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y1.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y1.append(0)\n",
    "print(\"Control vs Patients (Acute and Remission)\")\n",
    "plt.plot(steps,y, label = \"Control\")\n",
    "plt.plot(steps,y1, label = \"Patients (Acute and Remission)\")\n",
    "plt.xlabel(\"Overlap Coefficient Difference between 1st and 2nd highest Coefficent\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "steps = [0.0,0.01,0.02,0.03,0.04,0.05,0.06,0.07,0.08,0.09,0.1,0.11,0.12,0.13,0.14,0.15,0.16,0.17,0.18,0.19,0.2]\n",
    "controls_remission = []\n",
    "acute = []\n",
    "for participant,row in df_ngrams.iterrows():\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 0:\n",
    "        controls_remission.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 1:\n",
    "        acute.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "    if df_ngrams.loc[participant, \"Label\"] == 2:\n",
    "        controls_remission.append((round(df_ngrams.loc[participant, \"Overlap Difference\"],2),participant))\n",
    "\n",
    "y = []\n",
    "y1 = []\n",
    "\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in controls_remission:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "            elif df_ngrams.loc[tupl[1], \"Predicted Label\"] == 0:\n",
    "                lstcorre.append(tupl)\n",
    "            elif df_ngrams.loc[tupl[1], \"Predicted Label\"] == 2:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y.append(0)\n",
    "for number in steps:\n",
    "    lsttotal = []\n",
    "    lstcorre = []\n",
    "    graph = []\n",
    "    for tupl in acute:\n",
    "        if tupl[0] == number:\n",
    "            lsttotal.append(tupl)\n",
    "            if df_ngrams.loc[tupl[1], \"Predicted Label\"] == df_ngrams.loc[tupl[1], \"Label\"]:\n",
    "                lstcorre.append(tupl)\n",
    "    if len(lsttotal) != 0:\n",
    "        accuracy = len(lstcorre) / len(lsttotal)\n",
    "        y1.append(accuracy)\n",
    "    if len(lsttotal) == 0:\n",
    "        y1.append(0)\n",
    "print(\"Control and Remission vs Acute\")\n",
    "plt.plot(steps,y, label = \"Control and Remission\")\n",
    "plt.plot(steps,y1, label = \"Acute\")\n",
    "plt.xlabel(\"Overlap Coefficient Difference between 1st and 2nd highest Coefficent\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "def forming_graph(participant_ngrams, participant_group):\n",
    "    counting_ngrams = []\n",
    "    percentage_axis = []\n",
    "    seen_ngrams = []\n",
    "    for tupl in participant_group:\n",
    "        seen_ngrams.append(tupl)\n",
    "        coefficient = overlapp_coefficient(participant_ngrams, seen_ngrams)\n",
    "        percentage_axis.append(coefficient)\n",
    "    graph = pd.Series(percentage_axis)\n",
    "    return graph\n",
    "\n",
    "def plot_graph(participant_id, graph, graph1, **kwargs):\n",
    "    graph2 = kwargs.get(\"graph2\", None)\n",
    "    plt.plot(graph, label = \"(C) Controls\")\n",
    "    plt.plot(graph1, label = \"(PA) Acute Patients\")\n",
    "    plt.plot(graph2, label = \"(PR) Patients in Remission\")\n",
    "    plt.title(participant_id)\n",
    "    plt.xlabel('Ngrams seen')\n",
    "    plt.ylabel('Overlap Coefficient')\n",
    "    plt.legend()\n",
    "    plt.savefig(\"P352_overlap_plot\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ngrams = create_dataframe(\"both\",\"chars\", 8, \"three_groups\", participant_grouping = None)\n",
    "df_ngrams = df_ngrams.set_index(\"Participant\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Increase of overlap coefficient by incremental increase of \"seen\" ngrams for 8-char-ngrams"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "graph0 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Random 0\"])\n",
    "graph1 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Random 1\"])\n",
    "graph2 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Random 2\"])\n",
    "plot_graph(\"PR: 8-char-ngrams, both tasks, ngram samples of same size\", graph0, graph1, graph2 = graph2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "graph0 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Ngrams 0\"])\n",
    "graph1 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Ngrams 1\"])\n",
    "graph2 = forming_graph(df_ngrams.loc[\"P352\",\"Participant Ngrams\"], df_ngrams.loc[\"P352\",\"Ngrams 2\"])\n",
    "plot_graph(\"PR: 8-char-ngrams, both tasks, all ngrams\", graph0, graph1, graph2 = graph2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# graphs for the data from both tasks, character ngrams, ngram order 3 and three participant groups. \n",
    "# the graphs indicate how the overlap coefficient per participant changes while its getting fed the words/ngrams from the three test groups\n",
    "\n",
    "for participant, row in df_ngrams.iterrows():\n",
    "    participant_id = row[0]\n",
    "    participant_ngrams = df_ngrams.iloc[participant][\"Participant Ngrams\"]\n",
    "\n",
    "    graph0 = forming_graph(participant_ngrams, df_ngrams.iloc[participant][\"Ngrams 0\"])\n",
    "    graph1 = forming_graph(participant_ngrams, df_ngrams.iloc[participant][\"Ngrams 1\"])\n",
    "    graph2 = forming_graph(participant_ngrams, df_ngrams.iloc[participant][\"Ngrams 2\"])\n",
    "    plot_graph(participant_id, graph0, graph1, graph2 = graph2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Naive Bayes and Support Vector Machine"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Controls (0) versus Acute (1) versus Remission (2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Label</th>\n",
       "      <th>Taaloefening 1</th>\n",
       "      <th>Taaloefening 2</th>\n",
       "      <th>Text</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>P642</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C021</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C022</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C041</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C042</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Label                                     Taaloefening 1  \\\n",
       "ID                                                               \n",
       "P642      2                                                NaN   \n",
       "C021      0  Je neemt de nodige hoeveelheid water, deze gie...   \n",
       "C022      0                                                NaN   \n",
       "C041      0                                                NaN   \n",
       "C042      0  Je neemt een hoog bokaal, een filter. De filte...   \n",
       "\n",
       "                                         Taaloefening 2  \\\n",
       "ID                                                        \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...   \n",
       "C021                                                NaN   \n",
       "C022  De kok is een pizza aan het maken in zijn rest...   \n",
       "C041  Op de printjes zie je een pizza maker die een ...   \n",
       "C042                                                NaN   \n",
       "\n",
       "                                                   Text  \n",
       "ID                                                       \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...  \n",
       "C021  Je neemt de nodige hoeveelheid water, deze gie...  \n",
       "C022  De kok is een pizza aan het maken in zijn rest...  \n",
       "C041  Op de printjes zie je een pizza maker die een ...  \n",
       "C042  Je neemt een hoog bokaal, een filter. De filte...  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[7 1 0]\n",
      " [7 1 0]\n",
      " [1 0 0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.47      0.88      0.61         8\n",
      "           1       0.50      0.12      0.20         8\n",
      "           2       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.47        17\n",
      "   macro avg       0.32      0.33      0.27        17\n",
      "weighted avg       0.45      0.47      0.38        17\n",
      "\n",
      "Accuracy: 0.66 (+/- 0.15)\n",
      "K-Fold Cross Validation: Accuracy: 63.333% (+/- 35.355%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 61.905% (+/- 48.562%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 61.071% (+/- 11.232%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[7 1 0]\n",
      " [0 7 1]\n",
      " [0 1 0]]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 4 members, which is too few. The minimum number of members in any class cannot be less than n_splits=5.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 4 members, which is too few. The minimum number of members in any class cannot be less than n_splits=5.\n",
      "  % (min_groups, self.n_splits)), Warning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.88      0.93         8\n",
      "           1       0.78      0.88      0.82         8\n",
      "           2       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.82        17\n",
      "   macro avg       0.59      0.58      0.59        17\n",
      "weighted avg       0.84      0.82      0.83        17\n",
      "\n",
      "Accuracy: 0.67 (+/- 0.16)\n",
      "K-Fold Cross Validation: Accuracy: 67.667% (+/- 24.757%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 65.476% (+/- 47.545%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 67.143% (+/- 6.349%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[11  1  0]\n",
      " [ 3  1  0]\n",
      " [ 0  1  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.79      0.92      0.85        12\n",
      "           1       0.33      0.25      0.29         4\n",
      "           2       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.71        17\n",
      "   macro avg       0.37      0.39      0.38        17\n",
      "weighted avg       0.63      0.71      0.66        17\n",
      "\n",
      "Accuracy: 0.63 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 60.000% (+/- 37.118%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 57.143% (+/- 49.487%)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 4 members, which is too few. The minimum number of members in any class cannot be less than n_splits=5.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 4 members, which is too few. The minimum number of members in any class cannot be less than n_splits=5.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Repeated Random Test-Train Splits: Accuracy: 58.929% (+/- 8.338%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[9 3 0]\n",
      " [2 2 0]\n",
      " [1 0 0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.75      0.75        12\n",
      "           1       0.40      0.50      0.44         4\n",
      "           2       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.65        17\n",
      "   macro avg       0.38      0.42      0.40        17\n",
      "weighted avg       0.62      0.65      0.63        17\n",
      "\n",
      "Accuracy: 0.67 (+/- 0.22)\n",
      "K-Fold Cross Validation: Accuracy: 75.667% (+/- 23.912%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 73.810% (+/- 43.967%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 71.071% (+/- 6.071%)\n",
      "with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[21  0  0]\n",
      " [ 3  9  0]\n",
      " [ 0  1  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      1.00      0.93        21\n",
      "           1       0.90      0.75      0.82        12\n",
      "           2       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.88        34\n",
      "   macro avg       0.59      0.58      0.58        34\n",
      "weighted avg       0.86      0.88      0.87        34\n",
      "\n",
      "Accuracy: 0.87 (+/- 0.08)\n",
      "K-Fold Cross Validation: Accuracy: 85.619% (+/- 18.522%)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Leave-One-Out Cross Validation: Accuracy: 86.905% (+/- 33.735%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 76.607% (+/- 4.549%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[21  0  0]\n",
      " [ 0 12  0]\n",
      " [ 0  0  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        21\n",
      "           1       1.00      1.00      1.00        12\n",
      "           2       1.00      1.00      1.00         1\n",
      "\n",
      "    accuracy                           1.00        34\n",
      "   macro avg       1.00      1.00      1.00        34\n",
      "weighted avg       1.00      1.00      1.00        34\n",
      "\n",
      "Accuracy: 0.98 (+/- 0.03)\n",
      "K-Fold Cross Validation: Accuracy: 98.857% (+/- 3.876%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 98.810% (+/- 10.846%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 87.321% (+/- 6.209%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 1:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df.iterrows():\n",
    "    if isinstance(df.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])\n",
    "        lemmas.append(row[4])\n",
    "        y.append(row[0])\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 2: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[7 0 2]\n",
      " [1 0 1]\n",
      " [5 0 0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.54      0.78      0.64         9\n",
      "           1       0.00      0.00      0.00         2\n",
      "           2       0.00      0.00      0.00         5\n",
      "\n",
      "    accuracy                           0.44        16\n",
      "   macro avg       0.18      0.26      0.21        16\n",
      "weighted avg       0.30      0.44      0.36        16\n",
      "\n",
      "Accuracy: 0.52 (+/- 0.11)\n",
      "K-Fold Cross Validation: Accuracy: 52.667% (+/- 34.296%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 53.165% (+/- 49.900%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 54.815% (+/- 9.192%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[6 0 3]\n",
      " [1 0 1]\n",
      " [1 0 4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.67      0.71         9\n",
      "           1       0.00      0.00      0.00         2\n",
      "           2       0.50      0.80      0.62         5\n",
      "\n",
      "    accuracy                           0.62        16\n",
      "   macro avg       0.42      0.49      0.44        16\n",
      "weighted avg       0.58      0.62      0.59        16\n",
      "\n",
      "Accuracy: 0.61 (+/- 0.17)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-Fold Cross Validation: Accuracy: 61.333% (+/- 22.838%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 63.291% (+/- 48.201%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 62.963% (+/- 5.238%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[7 0 0]\n",
      " [0 1 0]\n",
      " [7 0 1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      1.00      0.67         7\n",
      "           1       1.00      1.00      1.00         1\n",
      "           2       1.00      0.12      0.22         8\n",
      "\n",
      "    accuracy                           0.56        16\n",
      "   macro avg       0.83      0.71      0.63        16\n",
      "weighted avg       0.78      0.56      0.47        16\n",
      "\n",
      "Accuracy: 0.56 (+/- 0.20)\n",
      "K-Fold Cross Validation: Accuracy: 57.667% (+/- 33.905%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 60.759% (+/- 48.829%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 55.556% (+/- 9.072%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[4 0 3]\n",
      " [0 0 1]\n",
      " [3 0 5]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.57      0.57      0.57         7\n",
      "           1       0.00      0.00      0.00         1\n",
      "           2       0.56      0.62      0.59         8\n",
      "\n",
      "    accuracy                           0.56        16\n",
      "   macro avg       0.38      0.40      0.39        16\n",
      "weighted avg       0.53      0.56      0.54        16\n",
      "\n",
      "Accuracy: 0.60 (+/- 0.20)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-Fold Cross Validation: Accuracy: 55.333% (+/- 26.654%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 58.228% (+/- 49.318%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 59.630% (+/- 6.718%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[18  0  1]\n",
      " [ 0  0  1]\n",
      " [ 8  0  4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.95      0.80        19\n",
      "           1       0.00      0.00      0.00         1\n",
      "           2       0.67      0.33      0.44        12\n",
      "\n",
      "    accuracy                           0.69        32\n",
      "   macro avg       0.45      0.43      0.41        32\n",
      "weighted avg       0.66      0.69      0.64        32\n",
      "\n",
      "Accuracy: 0.82 (+/- 0.08)\n",
      "K-Fold Cross Validation: Accuracy: 80.095% (+/- 21.048%)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Leave-One-Out Cross Validation: Accuracy: 81.646% (+/- 38.711%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 66.604% (+/- 4.548%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[19  0  0]\n",
      " [ 0  1  0]\n",
      " [ 0  0 12]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        19\n",
      "           1       1.00      1.00      1.00         1\n",
      "           2       1.00      1.00      1.00        12\n",
      "\n",
      "    accuracy                           1.00        32\n",
      "   macro avg       1.00      1.00      1.00        32\n",
      "weighted avg       1.00      1.00      1.00        32\n",
      "\n",
      "Accuracy: 0.99 (+/- 0.02)\n",
      "K-Fold Cross Validation: Accuracy: 98.857% (+/- 3.876%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 98.734% (+/- 11.179%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 83.208% (+/- 2.728%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 2:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df.iterrows():\n",
    "    if isinstance(df.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        lemmas.append(row[4])\n",
    "        y.append(row[0])\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1 & 2: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[16  3  0]\n",
      " [ 1  7  0]\n",
      " [ 4  0  2]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.84      0.80        19\n",
      "           1       0.70      0.88      0.78         8\n",
      "           2       1.00      0.33      0.50         6\n",
      "\n",
      "    accuracy                           0.76        33\n",
      "   macro avg       0.82      0.68      0.69        33\n",
      "weighted avg       0.79      0.76      0.74        33\n",
      "\n",
      "Accuracy: 0.58 (+/- 0.10)\n",
      "K-Fold Cross Validation: Accuracy: 51.238% (+/- 26.368%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 55.828% (+/- 49.659%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 57.222% (+/- 6.332%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[16  2  1]\n",
      " [ 1  6  1]\n",
      " [ 2  0  4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.84      0.84        19\n",
      "           1       0.75      0.75      0.75         8\n",
      "           2       0.67      0.67      0.67         6\n",
      "\n",
      "    accuracy                           0.79        33\n",
      "   macro avg       0.75      0.75      0.75        33\n",
      "weighted avg       0.79      0.79      0.79        33\n",
      "\n",
      "Accuracy: 0.65 (+/- 0.08)\n",
      "K-Fold Cross Validation: Accuracy: 63.429% (+/- 22.657%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 62.577% (+/- 48.392%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 64.815% (+/- 3.415%)\n",
      "\n",
      " with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[13  2  7]\n",
      " [ 3  6  0]\n",
      " [ 0  1  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.59      0.68        22\n",
      "           1       0.67      0.67      0.67         9\n",
      "           2       0.12      0.50      0.20         2\n",
      "\n",
      "    accuracy                           0.61        33\n",
      "   macro avg       0.53      0.59      0.52        33\n",
      "weighted avg       0.73      0.61      0.65        33\n",
      "\n",
      "Accuracy: 0.59 (+/- 0.20)\n",
      "K-Fold Cross Validation: Accuracy: 52.667% (+/- 26.362%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 58.282% (+/- 49.309%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 56.852% (+/- 7.222%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[15  2  5]\n",
      " [ 3  4  2]\n",
      " [ 0  1  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.83      0.68      0.75        22\n",
      "           1       0.57      0.44      0.50         9\n",
      "           2       0.12      0.50      0.20         2\n",
      "\n",
      "    accuracy                           0.61        33\n",
      "   macro avg       0.51      0.54      0.48        33\n",
      "weighted avg       0.72      0.61      0.65        33\n",
      "\n",
      "Accuracy: 0.64 (+/- 0.06)\n",
      "K-Fold Cross Validation: Accuracy: 63.333% (+/- 22.568%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 65.644% (+/- 47.490%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 67.407% (+/- 4.319%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[28  4  4]\n",
      " [ 4  9  0]\n",
      " [ 6  3  8]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.78      0.76        36\n",
      "           1       0.56      0.69      0.62        13\n",
      "           2       0.67      0.47      0.55        17\n",
      "\n",
      "    accuracy                           0.68        66\n",
      "   macro avg       0.66      0.65      0.64        66\n",
      "weighted avg       0.68      0.68      0.68        66\n",
      "\n",
      "Accuracy: 0.78 (+/- 0.15)\n",
      "K-Fold Cross Validation: Accuracy: 67.209% (+/- 18.700%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 70.245% (+/- 45.718%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 71.111% (+/- 3.967%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[32  2  2]\n",
      " [ 0 13  0]\n",
      " [ 2  2 13]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.89      0.91        36\n",
      "           1       0.76      1.00      0.87        13\n",
      "           2       0.87      0.76      0.81        17\n",
      "\n",
      "    accuracy                           0.88        66\n",
      "   macro avg       0.86      0.88      0.86        66\n",
      "weighted avg       0.89      0.88      0.88        66\n",
      "\n",
      "Accuracy: 0.98 (+/- 0.01)\n",
      "K-Fold Cross Validation: Accuracy: 98.154% (+/- 3.940%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 98.466% (+/- 12.289%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 84.259% (+/- 4.498%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 1 & 2:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df.iterrows():\n",
    "    lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "    if isinstance(df.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])\n",
    "        y.append(row[0])\n",
    "    if isinstance(df.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        y.append(row[0])\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "vectorizer = CountVectorizer()\n",
    "print(\"\\n\", \"with lemmas:\")\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, df[\"Label\"], test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,df[\"Label\"],X_train, y_train)\n",
    "classifierresults(X, df[\"Label\"], y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,df[\"Label\"],X_train, y_train)\n",
    "classifierresults(X, df[\"Label\"], y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Controls (0) versus Patients (1, Acute and Remission)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_controlpatients = df.copy()\n",
    "for participant,row in df_controlpatients.iterrows():\n",
    "    if row[0] == 2:\n",
    "        df_controlpatients.loc[participant, \"Label\"] = 1\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[ 6  0]\n",
      " [11  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.35      1.00      0.52         6\n",
      "           1       0.00      0.00      0.00        11\n",
      "\n",
      "    accuracy                           0.35        17\n",
      "   macro avg       0.18      0.50      0.26        17\n",
      "weighted avg       0.12      0.35      0.18        17\n",
      "\n",
      "Accuracy: 0.68 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 69.667% (+/- 31.700%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 69.048% (+/- 46.230%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 62.500% (+/- 10.504%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[5 1]\n",
      " [7 4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.42      0.83      0.56         6\n",
      "           1       0.80      0.36      0.50        11\n",
      "\n",
      "    accuracy                           0.53        17\n",
      "   macro avg       0.61      0.60      0.53        17\n",
      "weighted avg       0.66      0.53      0.52        17\n",
      "\n",
      "Accuracy: 0.69 (+/- 0.05)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-Fold Cross Validation: Accuracy: 67.667% (+/- 24.757%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 69.048% (+/- 46.230%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 69.643% (+/- 6.029%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[11  3]\n",
      " [ 2  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.79      0.81        14\n",
      "           1       0.25      0.33      0.29         3\n",
      "\n",
      "    accuracy                           0.71        17\n",
      "   macro avg       0.55      0.56      0.55        17\n",
      "weighted avg       0.74      0.71      0.72        17\n",
      "\n",
      "Accuracy: 0.62 (+/- 0.11)\n",
      "K-Fold Cross Validation: Accuracy: 58.333% (+/- 33.250%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 59.524% (+/- 49.085%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 62.857% (+/- 9.340%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[10  4]\n",
      " [ 1  2]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.91      0.71      0.80        14\n",
      "           1       0.33      0.67      0.44         3\n",
      "\n",
      "    accuracy                           0.71        17\n",
      "   macro avg       0.62      0.69      0.62        17\n",
      "weighted avg       0.81      0.71      0.74        17\n",
      "\n",
      "Accuracy: 0.73 (+/- 0.12)\n",
      "K-Fold Cross Validation: Accuracy: 74.333% (+/- 22.598%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 75.000% (+/- 43.301%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 72.857% (+/- 6.227%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[22  0]\n",
      " [ 1 11]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      1.00      0.98        22\n",
      "           1       1.00      0.92      0.96        12\n",
      "\n",
      "    accuracy                           0.97        34\n",
      "   macro avg       0.98      0.96      0.97        34\n",
      "weighted avg       0.97      0.97      0.97        34\n",
      "\n",
      "Accuracy: 0.89 (+/- 0.09)\n",
      "K-Fold Cross Validation: Accuracy: 87.429% (+/- 13.595%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 91.071% (+/- 28.516%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 80.357% (+/- 5.051%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[22  0]\n",
      " [ 0 12]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        22\n",
      "           1       1.00      1.00      1.00        12\n",
      "\n",
      "    accuracy                           1.00        34\n",
      "   macro avg       1.00      1.00      1.00        34\n",
      "weighted avg       1.00      1.00      1.00        34\n",
      "\n",
      "Accuracy: 1.00 (+/- 0.00)\n",
      "K-Fold Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 90.000% (+/- 5.428%)\n"
     ]
    }
   ],
   "source": [
    "# Control vs Patients\n",
    "print(\"Taaloefening 1:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_controlpatients.iterrows():\n",
    "    if isinstance(df_controlpatients.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "        y.append(row[0])\n",
    "        \n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Label</th>\n",
       "      <th>Taaloefening 1</th>\n",
       "      <th>Taaloefening 2</th>\n",
       "      <th>Text</th>\n",
       "      <th>Lemmas</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>P642</th>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "      <td>Terwijl de kok zijn deeg aan het voorbereiden ...</td>\n",
       "      <td>terwijl de kok zijn deeg aan_het voorbereiden ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C021</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt de nodige hoeveelheid water, deze gie...</td>\n",
       "      <td>je nemen de nodig hoeveelheid water , deze gie...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C022</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "      <td>De kok is een pizza aan het maken in zijn rest...</td>\n",
       "      <td>de kok zijn een pizza aan_het maken in zijn re...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C041</th>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "      <td>Op de printjes zie je een pizza maker die een ...</td>\n",
       "      <td>op de prin zien je een pizza maker die een dee...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C042</th>\n",
       "      <td>0</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Je neemt een hoog bokaal, een filter. De filte...</td>\n",
       "      <td>je nemen een hoog bokaal , een filter . de fil...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Label                                     Taaloefening 1  \\\n",
       "ID                                                               \n",
       "P642      1                                                NaN   \n",
       "C021      0  Je neemt de nodige hoeveelheid water, deze gie...   \n",
       "C022      0                                                NaN   \n",
       "C041      0                                                NaN   \n",
       "C042      0  Je neemt een hoog bokaal, een filter. De filte...   \n",
       "\n",
       "                                         Taaloefening 2  \\\n",
       "ID                                                        \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...   \n",
       "C021                                                NaN   \n",
       "C022  De kok is een pizza aan het maken in zijn rest...   \n",
       "C041  Op de printjes zie je een pizza maker die een ...   \n",
       "C042                                                NaN   \n",
       "\n",
       "                                                   Text  \\\n",
       "ID                                                        \n",
       "P642  Terwijl de kok zijn deeg aan het voorbereiden ...   \n",
       "C021  Je neemt de nodige hoeveelheid water, deze gie...   \n",
       "C022  De kok is een pizza aan het maken in zijn rest...   \n",
       "C041  Op de printjes zie je een pizza maker die een ...   \n",
       "C042  Je neemt een hoog bokaal, een filter. De filte...   \n",
       "\n",
       "                                                 Lemmas  \n",
       "ID                                                       \n",
       "P642  terwijl de kok zijn deeg aan_het voorbereiden ...  \n",
       "C021  je nemen de nodig hoeveelheid water , deze gie...  \n",
       "C022  de kok zijn een pizza aan_het maken in zijn re...  \n",
       "C041  op de prin zien je een pizza maker die een dee...  \n",
       "C042  je nemen een hoog bokaal , een filter . de fil...  "
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_controlpatients.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[9 4]\n",
      " [2 2]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.69      0.75        13\n",
      "           1       0.33      0.50      0.40         4\n",
      "\n",
      "    accuracy                           0.65        17\n",
      "   macro avg       0.58      0.60      0.58        17\n",
      "weighted avg       0.70      0.65      0.67        17\n",
      "\n",
      "Accuracy: 0.68 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 69.667% (+/- 31.700%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 69.048% (+/- 46.230%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 62.500% (+/- 10.504%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[8 5]\n",
      " [1 3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.89      0.62      0.73        13\n",
      "           1       0.38      0.75      0.50         4\n",
      "\n",
      "    accuracy                           0.65        17\n",
      "   macro avg       0.63      0.68      0.61        17\n",
      "weighted avg       0.77      0.65      0.67        17\n",
      "\n",
      "Accuracy: 0.69 (+/- 0.05)\n",
      "K-Fold Cross Validation: Accuracy: 67.667% (+/- 24.757%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 69.048% (+/- 46.230%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 69.643% (+/- 6.029%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[8 1]\n",
      " [5 3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.62      0.89      0.73         9\n",
      "           1       0.75      0.38      0.50         8\n",
      "\n",
      "    accuracy                           0.65        17\n",
      "   macro avg       0.68      0.63      0.61        17\n",
      "weighted avg       0.68      0.65      0.62        17\n",
      "\n",
      "Accuracy: 0.62 (+/- 0.11)\n",
      "K-Fold Cross Validation: Accuracy: 58.333% (+/- 33.250%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 59.524% (+/- 49.085%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 62.857% (+/- 9.340%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[6 3]\n",
      " [1 7]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.86      0.67      0.75         9\n",
      "           1       0.70      0.88      0.78         8\n",
      "\n",
      "    accuracy                           0.76        17\n",
      "   macro avg       0.78      0.77      0.76        17\n",
      "weighted avg       0.78      0.76      0.76        17\n",
      "\n",
      "Accuracy: 0.73 (+/- 0.12)\n",
      "K-Fold Cross Validation: Accuracy: 74.333% (+/- 22.598%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 75.000% (+/- 43.301%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 72.857% (+/- 6.227%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[22  0]\n",
      " [ 1 11]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      1.00      0.98        22\n",
      "           1       1.00      0.92      0.96        12\n",
      "\n",
      "    accuracy                           0.97        34\n",
      "   macro avg       0.98      0.96      0.97        34\n",
      "weighted avg       0.97      0.97      0.97        34\n",
      "\n",
      "Accuracy: 0.89 (+/- 0.09)\n",
      "K-Fold Cross Validation: Accuracy: 87.429% (+/- 13.595%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 91.071% (+/- 28.516%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 80.357% (+/- 5.051%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[22  0]\n",
      " [ 0 12]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00        22\n",
      "           1       1.00      1.00      1.00        12\n",
      "\n",
      "    accuracy                           1.00        34\n",
      "   macro avg       1.00      1.00      1.00        34\n",
      "weighted avg       1.00      1.00      1.00        34\n",
      "\n",
      "Accuracy: 1.00 (+/- 0.00)\n",
      "K-Fold Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 90.000% (+/- 5.428%)\n"
     ]
    }
   ],
   "source": [
    "# Control vs Patients\n",
    "print(\"Taaloefening 1:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_controlpatients.iterrows():\n",
    "    if isinstance(df_controlpatients.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "        y.append(row[0])\n",
    "        \n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 2: \n",
      "\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[6 1]\n",
      " [6 3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.86      0.63         7\n",
      "           1       0.75      0.33      0.46         9\n",
      "\n",
      "    accuracy                           0.56        16\n",
      "   macro avg       0.62      0.60      0.55        16\n",
      "weighted avg       0.64      0.56      0.54        16\n",
      "\n",
      "Accuracy: 0.56 (+/- 0.07)\n",
      "K-Fold Cross Validation: Accuracy: 56.667% (+/- 31.002%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 62.025% (+/- 48.532%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 58.148% (+/- 10.612%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[5 2]\n",
      " [3 6]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.62      0.71      0.67         7\n",
      "           1       0.75      0.67      0.71         9\n",
      "\n",
      "    accuracy                           0.69        16\n",
      "   macro avg       0.69      0.69      0.69        16\n",
      "weighted avg       0.70      0.69      0.69        16\n",
      "\n",
      "Accuracy: 0.71 (+/- 0.19)\n",
      "K-Fold Cross Validation: Accuracy: 65.667% (+/- 21.644%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 65.823% (+/- 47.430%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 66.667% (+/- 7.220%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[6 2]\n",
      " [7 1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.46      0.75      0.57         8\n",
      "           1       0.33      0.12      0.18         8\n",
      "\n",
      "    accuracy                           0.44        16\n",
      "   macro avg       0.40      0.44      0.38        16\n",
      "weighted avg       0.40      0.44      0.38        16\n",
      "\n",
      "Accuracy: 0.57 (+/- 0.16)\n",
      "K-Fold Cross Validation: Accuracy: 57.667% (+/- 26.554%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 62.025% (+/- 48.532%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 59.259% (+/- 10.075%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[6 2]\n",
      " [3 5]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.75      0.71         8\n",
      "           1       0.71      0.62      0.67         8\n",
      "\n",
      "    accuracy                           0.69        16\n",
      "   macro avg       0.69      0.69      0.69        16\n",
      "weighted avg       0.69      0.69      0.69        16\n",
      "\n",
      "Accuracy: 0.67 (+/- 0.20)\n",
      "K-Fold Cross Validation: Accuracy: 62.000% (+/- 22.494%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 63.291% (+/- 48.201%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 65.556% (+/- 6.839%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[20  0]\n",
      " [ 8  4]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.71      1.00      0.83        20\n",
      "           1       1.00      0.33      0.50        12\n",
      "\n",
      "    accuracy                           0.75        32\n",
      "   macro avg       0.86      0.67      0.67        32\n",
      "weighted avg       0.82      0.75      0.71        32\n",
      "\n",
      "Accuracy: 0.86 (+/- 0.03)\n",
      "K-Fold Cross Validation: Accuracy: 82.952% (+/- 13.362%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 86.709% (+/- 33.948%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 71.887% (+/- 3.624%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[18  2]\n",
      " [ 2 10]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.90      0.90      0.90        20\n",
      "           1       0.83      0.83      0.83        12\n",
      "\n",
      "    accuracy                           0.88        32\n",
      "   macro avg       0.87      0.87      0.87        32\n",
      "weighted avg       0.88      0.88      0.88        32\n",
      "\n",
      "Accuracy: 1.00 (+/- 0.00)\n",
      "K-Fold Cross Validation: Accuracy: 98.857% (+/- 3.876%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 98.734% (+/- 11.179%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 85.283% (+/- 3.560%)\n"
     ]
    }
   ],
   "source": [
    "# Control vs Patients\n",
    "print(\"Taaloefening 2:\", \"\\n\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_controlpatients.iterrows():\n",
    "    if isinstance(df_controlpatients.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "        y.append(row[0])\n",
    "        \n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1 & 2:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[12  2]\n",
      " [18  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.40      0.86      0.55        14\n",
      "           1       0.33      0.05      0.09        19\n",
      "\n",
      "    accuracy                           0.39        33\n",
      "   macro avg       0.37      0.45      0.32        33\n",
      "weighted avg       0.36      0.39      0.28        33\n",
      "\n",
      "Accuracy: 0.56 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 58.952% (+/- 30.958%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 60.123% (+/- 48.965%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 59.815% (+/- 5.620%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[9 9]\n",
      " [8 7]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      0.50      0.51        18\n",
      "           1       0.44      0.47      0.45        15\n",
      "\n",
      "    accuracy                           0.48        33\n",
      "   macro avg       0.48      0.48      0.48        33\n",
      "weighted avg       0.49      0.48      0.49        33\n",
      "\n",
      "Accuracy: 0.70 (+/- 0.06)\n",
      "K-Fold Cross Validation: Accuracy: 70.762% (+/- 12.335%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 73.620% (+/- 44.069%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 70.370% (+/- 3.208%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[14  1]\n",
      " [13  5]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.93      0.67        15\n",
      "           1       0.83      0.28      0.42        18\n",
      "\n",
      "    accuracy                           0.58        33\n",
      "   macro avg       0.68      0.61      0.54        33\n",
      "weighted avg       0.69      0.58      0.53        33\n",
      "\n",
      "Accuracy: 0.57 (+/- 0.08)\n",
      "K-Fold Cross Validation: Accuracy: 57.810% (+/- 31.077%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 61.350% (+/- 48.695%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 58.519% (+/- 6.738%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[10  5]\n",
      " [ 5 13]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.67      0.67      0.67        15\n",
      "           1       0.72      0.72      0.72        18\n",
      "\n",
      "    accuracy                           0.70        33\n",
      "   macro avg       0.69      0.69      0.69        33\n",
      "weighted avg       0.70      0.70      0.70        33\n",
      "\n",
      "Accuracy: 0.71 (+/- 0.12)\n",
      "K-Fold Cross Validation: Accuracy: 70.095% (+/- 15.284%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 71.779% (+/- 45.007%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 71.667% (+/- 3.318%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[30  5]\n",
      " [11 20]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      0.86      0.79        35\n",
      "           1       0.80      0.65      0.71        31\n",
      "\n",
      "    accuracy                           0.76        66\n",
      "   macro avg       0.77      0.75      0.75        66\n",
      "weighted avg       0.76      0.76      0.75        66\n",
      "\n",
      "Accuracy: 0.84 (+/- 0.05)\n",
      "K-Fold Cross Validation: Accuracy: 80.352% (+/- 16.736%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 83.742% (+/- 36.898%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 74.167% (+/- 4.855%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[34  1]\n",
      " [ 8 23]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      0.97      0.88        35\n",
      "           1       0.96      0.74      0.84        31\n",
      "\n",
      "    accuracy                           0.86        66\n",
      "   macro avg       0.88      0.86      0.86        66\n",
      "weighted avg       0.88      0.86      0.86        66\n",
      "\n",
      "Accuracy: 0.98 (+/- 0.02)\n",
      "K-Fold Cross Validation: Accuracy: 99.077% (+/- 2.500%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 99.080% (+/- 9.549%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 88.796% (+/- 4.129%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 1 & 2:\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_controlpatients.iterrows():\n",
    "    lemmas.append(df_controlpatients.loc[participant, \"Lemmas\"])\n",
    "    if isinstance(df_controlpatients.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])         \n",
    "        y.append(row[0])\n",
    "    if isinstance(df_controlpatients.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        y.append(row[0])\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X, y, X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X, y, X_train, y_train)\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Controls (0) and Patients in Remission versus Acute Patients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_conrem_acute = df.copy()\n",
    "for participant,row in df_conrem_acute.iterrows():\n",
    "    if row[0] == 2:\n",
    "        df_conrem_acute.loc[participant, \"Label\"] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[9 0]\n",
      " [8 0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      1.00      0.69         9\n",
      "           1       0.00      0.00      0.00         8\n",
      "\n",
      "    accuracy                           0.53        17\n",
      "   macro avg       0.26      0.50      0.35        17\n",
      "weighted avg       0.28      0.53      0.37        17\n",
      "\n",
      "Accuracy: 0.69 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 66.667% (+/- 31.798%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 63.095% (+/- 48.255%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 65.357% (+/- 8.755%)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[9 3]\n",
      " [2 3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.75      0.78        12\n",
      "           1       0.50      0.60      0.55         5\n",
      "\n",
      "    accuracy                           0.71        17\n",
      "   macro avg       0.66      0.68      0.66        17\n",
      "weighted avg       0.72      0.71      0.71        17\n",
      "\n",
      "Accuracy: 0.75 (+/- 0.16)\n",
      "K-Fold Cross Validation: Accuracy: 72.000% (+/- 22.716%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 70.238% (+/- 45.721%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 72.857% (+/- 5.803%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[9 1]\n",
      " [6 1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.60      0.90      0.72        10\n",
      "           1       0.50      0.14      0.22         7\n",
      "\n",
      "    accuracy                           0.59        17\n",
      "   macro avg       0.55      0.52      0.47        17\n",
      "weighted avg       0.56      0.59      0.52        17\n",
      "\n",
      "Accuracy: 0.67 (+/- 0.09)\n",
      "K-Fold Cross Validation: Accuracy: 64.667% (+/- 34.098%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 60.714% (+/- 48.839%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 63.929% (+/- 7.044%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[6 4]\n",
      " [1 6]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.86      0.60      0.71        10\n",
      "           1       0.60      0.86      0.71         7\n",
      "\n",
      "    accuracy                           0.71        17\n",
      "   macro avg       0.73      0.73      0.71        17\n",
      "weighted avg       0.75      0.71      0.71        17\n",
      "\n",
      "Accuracy: 0.73 (+/- 0.18)\n",
      "K-Fold Cross Validation: Accuracy: 78.333% (+/- 22.973%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 76.190% (+/- 42.592%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 74.286% (+/- 4.165%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[19  0]\n",
      " [ 7  8]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      1.00      0.84        19\n",
      "           1       1.00      0.53      0.70        15\n",
      "\n",
      "    accuracy                           0.79        34\n",
      "   macro avg       0.87      0.77      0.77        34\n",
      "weighted avg       0.85      0.79      0.78        34\n",
      "\n",
      "Accuracy: 0.91 (+/- 0.07)\n",
      "K-Fold Cross Validation: Accuracy: 90.190% (+/- 12.641%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 88.690% (+/- 31.671%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 80.536% (+/- 5.324%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[15  4]\n",
      " [ 0 15]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.79      0.88        19\n",
      "           1       0.79      1.00      0.88        15\n",
      "\n",
      "    accuracy                           0.88        34\n",
      "   macro avg       0.89      0.89      0.88        34\n",
      "weighted avg       0.91      0.88      0.88        34\n",
      "\n",
      "Accuracy: 1.00 (+/- 0.00)\n",
      "K-Fold Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 100.000% (+/- 0.000%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 88.750% (+/- 7.145%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 1:\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_conrem_acute.iterrows():\n",
    "    if isinstance(df_conrem_acute.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])         \n",
    "        y.append(row[0])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X, y, X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X, y, X_train, y_train)\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening2:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[14  0]\n",
      " [ 1  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      1.00      0.97        14\n",
      "           1       1.00      0.50      0.67         2\n",
      "\n",
      "    accuracy                           0.94        16\n",
      "   macro avg       0.97      0.75      0.82        16\n",
      "weighted avg       0.94      0.94      0.93        16\n",
      "\n",
      "Accuracy: 0.95 (+/- 0.05)\n",
      "K-Fold Cross Validation: Accuracy: 93.667% (+/- 15.861%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 93.671% (+/- 24.349%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 94.444% (+/- 3.795%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[13  0]\n",
      " [ 3  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.81      1.00      0.90        13\n",
      "           1       0.00      0.00      0.00         3\n",
      "\n",
      "    accuracy                           0.81        16\n",
      "   macro avg       0.41      0.50      0.45        16\n",
      "weighted avg       0.66      0.81      0.73        16\n",
      "\n",
      "Accuracy: 0.94 (+/- 0.00)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-Fold Cross Validation: Accuracy: 92.333% (+/- 16.653%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 93.671% (+/- 24.349%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 92.963% (+/- 3.867%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[15  0]\n",
      " [ 1  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      1.00      0.97        15\n",
      "           1       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.94        16\n",
      "   macro avg       0.47      0.50      0.48        16\n",
      "weighted avg       0.88      0.94      0.91        16\n",
      "\n",
      "Accuracy: 0.95 (+/- 0.05)\n",
      "K-Fold Cross Validation: Accuracy: 95.000% (+/- 14.907%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 93.671% (+/- 24.349%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 94.444% (+/- 3.795%)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/lisabecker/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[14  1]\n",
      " [ 1  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.93      0.93        15\n",
      "           1       0.00      0.00      0.00         1\n",
      "\n",
      "    accuracy                           0.88        16\n",
      "   macro avg       0.47      0.47      0.47        16\n",
      "weighted avg       0.88      0.88      0.88        16\n",
      "\n",
      "Accuracy: 0.92 (+/- 0.05)\n",
      "K-Fold Cross Validation: Accuracy: 93.667% (+/- 15.861%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 93.671% (+/- 24.349%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 92.222% (+/- 3.867%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[27  0]\n",
      " [ 4  1]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.87      1.00      0.93        27\n",
      "           1       1.00      0.20      0.33         5\n",
      "\n",
      "    accuracy                           0.88        32\n",
      "   macro avg       0.94      0.60      0.63        32\n",
      "weighted avg       0.89      0.88      0.84        32\n",
      "\n",
      "Accuracy: 0.95 (+/- 0.03)\n",
      "K-Fold Cross Validation: Accuracy: 95.048% (+/- 12.324%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 94.937% (+/- 21.925%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 94.528% (+/- 2.594%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[27  0]\n",
      " [ 2  3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      1.00      0.96        27\n",
      "           1       1.00      0.60      0.75         5\n",
      "\n",
      "    accuracy                           0.94        32\n",
      "   macro avg       0.97      0.80      0.86        32\n",
      "weighted avg       0.94      0.94      0.93        32\n",
      "\n",
      "Accuracy: 0.99 (+/- 0.02)\n",
      "K-Fold Cross Validation: Accuracy: 99.429% (+/- 2.799%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 99.367% (+/- 7.930%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 96.038% (+/- 1.016%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening2:\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_conrem_acute.iterrows():\n",
    "    if isinstance(df_conrem_acute.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        y.append(row[0])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X, y, X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X, y, X_train, y_train)\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening2:\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'df_conrem_acute' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-4-a155e3b15a2b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mlemmas\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0;32mfor\u001b[0m \u001b[0mparticipant\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrow\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdf_conrem_acute\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0miterrows\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf_conrem_acute\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mparticipant\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Taaloefening 2\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m         \u001b[0mcorpus\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrow\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'df_conrem_acute' is not defined"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening2:\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_conrem_acute.iterrows():\n",
    "    if isinstance(df_conrem_acute.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        y.append(row[0])\n",
    "        lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X, y, X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X, y, X_train, y_train)\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Taaloefening 1 & 2:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[22  4]\n",
      " [ 4  3]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.85      0.85        26\n",
      "           1       0.43      0.43      0.43         7\n",
      "\n",
      "    accuracy                           0.76        33\n",
      "   macro avg       0.64      0.64      0.64        33\n",
      "weighted avg       0.76      0.76      0.76        33\n",
      "\n",
      "Accuracy: 0.78 (+/- 0.10)\n",
      "K-Fold Cross Validation: Accuracy: 77.333% (+/- 15.210%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 76.687% (+/- 42.282%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 75.556% (+/- 6.075%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[23  5]\n",
      " [ 5  0]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.82      0.82        28\n",
      "           1       0.00      0.00      0.00         5\n",
      "\n",
      "    accuracy                           0.70        33\n",
      "   macro avg       0.41      0.41      0.41        33\n",
      "weighted avg       0.70      0.70      0.70        33\n",
      "\n",
      "Accuracy: 0.79 (+/- 0.04)\n",
      "K-Fold Cross Validation: Accuracy: 80.286% (+/- 15.482%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 79.141% (+/- 40.630%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 82.037% (+/- 3.104%)\n",
      "with lemmas:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[23  1]\n",
      " [ 7  2]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      0.96      0.85        24\n",
      "           1       0.67      0.22      0.33         9\n",
      "\n",
      "    accuracy                           0.76        33\n",
      "   macro avg       0.72      0.59      0.59        33\n",
      "weighted avg       0.74      0.76      0.71        33\n",
      "\n",
      "Accuracy: 0.75 (+/- 0.09)\n",
      "K-Fold Cross Validation: Accuracy: 75.429% (+/- 18.781%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 76.074% (+/- 42.663%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 75.370% (+/- 4.615%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[23  1]\n",
      " [ 4  5]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.96      0.90        24\n",
      "           1       0.83      0.56      0.67         9\n",
      "\n",
      "    accuracy                           0.85        33\n",
      "   macro avg       0.84      0.76      0.78        33\n",
      "weighted avg       0.85      0.85      0.84        33\n",
      "\n",
      "Accuracy: 0.77 (+/- 0.09)\n",
      "K-Fold Cross Validation: Accuracy: 79.238% (+/- 18.790%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 79.141% (+/- 40.630%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 80.741% (+/- 3.902%)\n",
      "\n",
      " with BOTH:\n",
      "---------- Naive Bayes ----------\n",
      "Confusion Matrix: \n",
      " [[42  9]\n",
      " [ 0 15]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.82      0.90        51\n",
      "           1       0.62      1.00      0.77        15\n",
      "\n",
      "    accuracy                           0.86        66\n",
      "   macro avg       0.81      0.91      0.84        66\n",
      "weighted avg       0.91      0.86      0.87        66\n",
      "\n",
      "Accuracy: 0.83 (+/- 0.13)\n",
      "K-Fold Cross Validation: Accuracy: 79.495% (+/- 14.092%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 80.675% (+/- 39.485%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 83.056% (+/- 2.554%)\n",
      "---------- Support Vector Machine ---------- \n",
      "\n",
      "Confusion Matrix: \n",
      " [[49  2]\n",
      " [ 0 15]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.96      0.98        51\n",
      "           1       0.88      1.00      0.94        15\n",
      "\n",
      "    accuracy                           0.97        66\n",
      "   macro avg       0.94      0.98      0.96        66\n",
      "weighted avg       0.97      0.97      0.97        66\n",
      "\n",
      "Accuracy: 0.98 (+/- 0.03)\n",
      "K-Fold Cross Validation: Accuracy: 99.077% (+/- 2.500%)\n",
      "Leave-One-Out Cross Validation: Accuracy: 99.080% (+/- 9.549%)\n",
      "Repeated Random Test-Train Splits: Accuracy: 91.481% (+/- 2.260%)\n"
     ]
    }
   ],
   "source": [
    "print(\"Taaloefening 1 & 2:\")\n",
    "corpus = []\n",
    "lemmas = []\n",
    "y = []\n",
    "for participant, row in df_conrem_acute.iterrows():\n",
    "    lemmas.append(df.loc[participant, \"Lemmas\"])\n",
    "    if isinstance(df_conrem_acute.loc[participant, \"Taaloefening 1\"], float) is not True:\n",
    "        corpus.append(row[1])         \n",
    "        y.append(row[0])\n",
    "    if isinstance(df_conrem_acute.loc[participant, \"Taaloefening 2\"], float) is not True:\n",
    "        corpus.append(row[2])\n",
    "        y.append(row[0])\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X, y, X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X, y, X_train, y_train)\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"with lemmas:\")\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(lemmas)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "print(\"\\n\", \"with BOTH:\")\n",
    "corpus += lemmas\n",
    "y += y\n",
    "vectorizer = CountVectorizer()\n",
    "X = vectorizer.fit_transform(corpus)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)\n",
    "model, y_pred, scores = naivebayes(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)\n",
    "model, y_pred, scores = supportvectormachine(X,y,X_train, y_train)\n",
    "classifierresults(X, y, y_test, y_pred, scores, model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1844\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# one of the trials to extract the most important features; the problem was that the classifiers were to different\n",
    "# to find a way to extract features in a comparable way\n",
    "\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.svm import LinearSVC\n",
    "import matplotlib.pyplot as plt\n",
    "def plot_coefficients(classifier, feature_names, top_features=10):\n",
    "    coef = classifier.coef_.ravel()\n",
    "    top_positive_coefficients = np.argsort(coef)[-top_features:]\n",
    "    top_negative_coefficients = np.argsort(coef)[:top_features]\n",
    "    top_coefficients = np.hstack([top_negative_coefficients, top_positive_coefficients])\n",
    " # create plot\n",
    "    plt.figure(figsize=(15, 5))\n",
    "    colors = [\"red\" if c < 0 else \"blue\" for c in coef[top_coefficients]]\n",
    "    plt.bar(np.arange(2 * top_features), coef[top_coefficients], color=colors)\n",
    "    feature_names = np.array(feature_names)\n",
    "    plt.xticks(np.arange(1, 1 + 2 * top_features), feature_names[top_coefficients], rotation=60, ha=\"right\")\n",
    "    plt.show()\n",
    "cv = CountVectorizer()\n",
    "cv.fit(corpus)\n",
    "print(len(cv.vocabulary_))\n",
    "#print(cv.get_feature_names())\n",
    "X_train = cv.transform(corpus)\n",
    "\n",
    "svm = LinearSVC()\n",
    "svm.fit(X_train, y)\n",
    "plot_coefficients(svm, cv.get_feature_names())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "metadata": {
     "collapsed": false
    },
    "source": []
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}