RF_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest - Connie"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import helper_RF as l\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import SCORERS\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore') \n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import MinMaxScaler       # scaling data\n",
    "from sklearn.neighbors import KNeighborsRegressor    # regressor\n",
    "from sklearn.model_selection import GridSearchCV     # for grid search\n",
    "from sklearn.pipeline import make_pipeline           # for making pipelines\n",
    "from sklearn.feature_selection import SelectKBest\n",
    "from sklearn.feature_selection import f_regression\n",
    "from sklearn.feature_selection import mutual_info_regression\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#copy from the main file\n",
    "import math\n",
    "from datetime import datetime\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns # for visualiation\n",
    "import matplotlib.pyplot as plt # plotting\n",
    "import matplotlib\n",
    "matplotlib.style.use('ggplot')\n",
    "%matplotlib inline\n",
    "import altair as alt\n",
    "alt.renderers.enable('notebook')\n",
    "\n",
    "df = pd.read_csv('./facebook_with_reactions.csv')\n",
    "outcomes = df[['Rating']]\n",
    "\n",
    "## drop useless columns \n",
    "df = df.drop(columns=['Debate', 'status_link', 'permalink_url', 'Post URL', \n",
    "                      'status_message', 'link_name', 'share_count', 'Unnamed: 0', 'account_id',\n",
    "                      'status_id', 'status_type', 'status_published'])\n",
    "\n",
    "## remove spaces in columns\n",
    "df.columns = [c.replace(' ', '_') for c in df.columns]\n",
    "\n",
    "## Create copy of data for modeling\n",
    "model_data = df.copy()\n",
    "\n",
    "## change Rating to numeric for classification\n",
    "model_data.Rating = model_data.Rating.replace(['mostly false', 'no factual content', 'mixture of true and false', 'mostly true'], [0,1,2,3]).astype(int)\n",
    "model_data = model_data.drop(columns=['post_id', 'reaction_count'])\n",
    "#model_data = model_data.drop(columns=['Date_Published_2016-09-20','Date_Published_2016-09-21',\n",
    "#                                     'Date_Published_2016-09-22','Date_Published_2016-09-23',\n",
    "#                                     'Date_Published_2016-09-26','Date_Published_2016-09-27'])\n",
    "## make dummy columns (make categories numeric)\n",
    "model_data = pd.get_dummies(model_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#save model_data as csv file\n",
    "model_data.to_csv(\"./model_data.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#transfer the model data to one with attributes\n",
    "# and other with only the response (target), which is \"Rating\"\n",
    "df_target=model_data['Rating']\n",
    "df_dropped=model_data\n",
    "df_dropped=df_dropped.drop(columns=['Rating'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Null Values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#propagate non-null values forward, given that we do have null values in the dataset\n",
    "df_dropped.fillna(method='ffill',inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Select Variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Rating',\n",
       " 'comment_count',\n",
       " 'num_reactions',\n",
       " 'num_comments',\n",
       " 'num_shares',\n",
       " 'num_likes',\n",
       " 'num_loves',\n",
       " 'num_wows',\n",
       " 'num_hahas',\n",
       " 'num_sads',\n",
       " 'num_angrys',\n",
       " 'Category_left',\n",
       " 'Category_mainstream',\n",
       " 'Category_right',\n",
       " 'Page_ABC News Politics',\n",
       " 'Page_Addicting Info',\n",
       " 'Page_CNN Politics',\n",
       " 'Page_Eagle Rising',\n",
       " 'Page_Occupy Democrats',\n",
       " 'Page_Politico',\n",
       " 'Page_Right Wing News',\n",
       " 'Page_The Other 98%',\n",
       " 'Date_Published_2016-09-19',\n",
       " 'Date_Published_2016-09-20',\n",
       " 'Date_Published_2016-09-21',\n",
       " 'Date_Published_2016-09-22',\n",
       " 'Date_Published_2016-09-23',\n",
       " 'Date_Published_2016-09-26',\n",
       " 'Date_Published_2016-09-27',\n",
       " 'Post_Type_link',\n",
       " 'Post_Type_photo',\n",
       " 'Post_Type_text',\n",
       " 'Post_Type_video']"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#this gives a complet list of variables\n",
    "list(model_data.columns.values) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['num_comments', 'num_reactions', 'num_likes', 'comment_count', 'num_shares']\n"
     ]
    }
   ],
   "source": [
    "#this function gives the selected features, using Extra-Tree classifier\n",
    "df_selected_1=l.feature_selection(df_dropped,df_target,model_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>num_comments</th>\n",
       "      <th>num_reactions</th>\n",
       "      <th>num_likes</th>\n",
       "      <th>comment_count</th>\n",
       "      <th>num_shares</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2166.000000</td>\n",
       "      <td>2166.000000</td>\n",
       "      <td>2166.000000</td>\n",
       "      <td>2166.000000</td>\n",
       "      <td>2166.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>452.082179</td>\n",
       "      <td>5296.478763</td>\n",
       "      <td>4067.789935</td>\n",
       "      <td>441.455217</td>\n",
       "      <td>3834.459834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1510.787707</td>\n",
       "      <td>18159.453476</td>\n",
       "      <td>14940.861709</td>\n",
       "      <td>1324.224921</td>\n",
       "      <td>22345.925304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>29.000000</td>\n",
       "      <td>115.000000</td>\n",
       "      <td>78.000000</td>\n",
       "      <td>34.000000</td>\n",
       "      <td>18.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>118.500000</td>\n",
       "      <td>480.000000</td>\n",
       "      <td>297.000000</td>\n",
       "      <td>126.500000</td>\n",
       "      <td>76.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>374.000000</td>\n",
       "      <td>2211.500000</td>\n",
       "      <td>1466.250000</td>\n",
       "      <td>386.000000</td>\n",
       "      <td>556.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>35799.000000</td>\n",
       "      <td>330144.000000</td>\n",
       "      <td>280694.000000</td>\n",
       "      <td>32419.000000</td>\n",
       "      <td>401465.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       num_comments  num_reactions      num_likes  comment_count  \\\n",
       "count   2166.000000    2166.000000    2166.000000    2166.000000   \n",
       "mean     452.082179    5296.478763    4067.789935     441.455217   \n",
       "std     1510.787707   18159.453476   14940.861709    1324.224921   \n",
       "min        0.000000       0.000000       0.000000       0.000000   \n",
       "25%       29.000000     115.000000      78.000000      34.000000   \n",
       "50%      118.500000     480.000000     297.000000     126.500000   \n",
       "75%      374.000000    2211.500000    1466.250000     386.000000   \n",
       "max    35799.000000  330144.000000  280694.000000   32419.000000   \n",
       "\n",
       "          num_shares  \n",
       "count    2166.000000  \n",
       "mean     3834.459834  \n",
       "std     22345.925304  \n",
       "min         0.000000  \n",
       "25%        18.000000  \n",
       "50%        76.000000  \n",
       "75%       556.000000  \n",
       "max    401465.000000  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#take a look at the selected df\n",
    "df_selected_1.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Split Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Split Data as training and test, with size of 0.8 and 0.2\n",
    "\n",
    "#df_selected=selecting the ideal cols\n",
    "\n",
    "#df_target=with the outcome\n",
    "\n",
    "train_features,test_features,train_outcome,test_outcome=train_test_split(\n",
    "    df_selected_1,\n",
    "    df_target,\n",
    "    test_size=0.20,\n",
    "    random_state=11,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grid Search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'max_features': 'sqrt', 'min_samples_split': 10, 'n_estimators': 7}\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"
    }
   ],
   "source": [
    "#dropped is the dataframe with numeric values\n",
    "#selected is the one with all the selected variables\n",
    "\n",
    "l.model_selection_random_forest(df_selected_1,df_target,train_features,train_outcome)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Fitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap='true',\n",
       "            class_weight={0: 1, 1: 50, 2: 50, 3: 50}, criterion='gini',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=75,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=8, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#fill in all the best parameters\n",
    "w=50\n",
    "best_model_1=RandomForestClassifier(bootstrap='true',max_features='auto',min_samples_split=75,n_estimators=8,class_weight={0: 1, 1: w,2:w,3:w})\n",
    "best_model_1.fit(train_features,train_outcome) #79.7\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Graphs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1121489b0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.scatter(best_model_1.predict(test_features),test_outcome)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cross Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7601479915433403"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#cross validation by our best model\n",
    "l.cross_val(test_features,test_outcome,best_model_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.804147465437788"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#see the accuracy score\n",
    "accuracy_score(test_outcome, best_model_1.predict(test_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3, 1, 3, 3, 1, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       1, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3, 3, 1, 3, 2, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 1,\n",
       "       3, 3, 1, 1, 3, 3, 3, 3, 1, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3,\n",
       "       3, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 1, 3, 3, 3,\n",
       "       2, 3, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3, 3, 3, 3, 3, 3, 3, 1,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,\n",
       "       3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 3])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#see the predictions\n",
    "best_model_1.predict(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  0,   1,   0,  12],\n",
       "       [  0,  20,   2,  27],\n",
       "       [  0,   2,   0,  36],\n",
       "       [  0,   5,   0, 329]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#the result of confusion matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "y_pred=best_model_1.predict(test_features)\n",
    "confusion_matrix(test_outcome, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7448124041357714"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#the F1 score, \n",
    "from sklearn.metrics import f1_score\n",
    "f1_score(test_outcome, y_pred,average=\"weighted\")"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}