performance_prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "suitable-night",
   "metadata": {},
   "source": [
    "# Предварительный анализ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "employed-murray",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import plotly.express as px\n",
    "\n",
    "import statsmodels.api as sm \n",
    "from statsmodels.formula.api import glm\n",
    "\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from IPython.display import HTML, SVG, display\n",
    "from graphviz import Source\n",
    "\n",
    "\n",
    "def print_roc(y_test, predictions):\n",
    "    fpr, tpr, thresholds = roc_curve(y_test, predictions)\n",
    "    roc_auc = auc(fpr, tpr)\n",
    "    plt.figure()\n",
    "    lw = 2\n",
    "    plt.plot(fpr, tpr, color='darkorange',\n",
    "             lw=lw, label='ROC curve (area = %0.2f)' % roc_auc)\n",
    "    plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('ROC curve')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "employed-garlic",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/StudentsPerformance.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "closing-wayne",
   "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>gender</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>parental_level_of_education</th>\n",
       "      <th>lunch</th>\n",
       "      <th>test_preparation_course</th>\n",
       "      <th>have_no_problem</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>bachelor's degree</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>female</td>\n",
       "      <td>C</td>\n",
       "      <td>some college</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>master's degree</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>associate's degree</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>some college</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>associate's degree</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>some college</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>some college</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>male</td>\n",
       "      <td>D</td>\n",
       "      <td>high school</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>high school</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gender ethnicity parental_level_of_education  lunch  \\\n",
       "0  female         B           bachelor's degree      1   \n",
       "1  female         C                some college      1   \n",
       "2  female         B             master's degree      1   \n",
       "3    male         A          associate's degree      0   \n",
       "4    male         C                some college      1   \n",
       "5  female         B          associate's degree      1   \n",
       "6  female         B                some college      1   \n",
       "7    male         B                some college      0   \n",
       "8    male         D                 high school      0   \n",
       "9  female         B                 high school      0   \n",
       "\n",
       "   test_preparation_course  have_no_problem  \n",
       "0                        0                1  \n",
       "1                        1                1  \n",
       "2                        0                1  \n",
       "3                        0                0  \n",
       "4                        0                1  \n",
       "5                        0                1  \n",
       "6                        1                1  \n",
       "7                        0                0  \n",
       "8                        1                1  \n",
       "9                        0                0  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "straight-uncertainty",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='test_preparation_course', ylabel='count'>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Отрисовка гистограмм столбцов исходных данных\n",
    "fig, ax = plt.subplots(3, 2)\n",
    "fig.tight_layout()\n",
    "ax[2, 0].axis('off')\n",
    "\n",
    "sns.histplot(data=df, x='gender', ax=ax[0, 0])\n",
    "sns.histplot(data=df, x='ethnicity', ax=ax[0, 1])\n",
    "sns.histplot(data=df, x='parental_level_of_education', ax=ax[1, 0])\n",
    "\n",
    "ax[1, 0].xaxis.set_visible(True)\n",
    "for tick in ax[1, 0].get_xticklabels():\n",
    "    tick.set_rotation(90)\n",
    "\n",
    "sns.countplot(data=df, x='lunch', ax=ax[1, 1])\n",
    "sns.countplot(data=df, x='test_preparation_course', ax=ax[2, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "annual-sharp",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Построение точечного 3D графика для учебных предметов\n",
    "# fig = px.scatter_3d(df, x='math_score', y='rus_score', z='literature_score',\n",
    "#               color='have_no_problem', height=600)\n",
    "# fig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "interesting-single",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Построение попарных точечных графиков и гистограмм для отдельных учебных предметов\n",
    "# ax = sns.pairplot(df[['math_score', 'literature_score', 'rus_score']])\n",
    "# ax"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "instrumental-integral",
   "metadata": {},
   "source": [
    "# Создание и использование моделей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "unknown-pierce",
   "metadata": {},
   "outputs": [],
   "source": [
    "features = np.array(['gender', 'ethnicity', 'lunch', 'test_preparation_course', 'parental_level_of_education'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "weird-norman",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df[features]\n",
    "y = df['have_no_problem']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "caroline-yorkshire",
   "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>gender</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>lunch</th>\n",
       "      <th>test_preparation_course</th>\n",
       "      <th>parental_level_of_education</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>bachelor's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>female</td>\n",
       "      <td>C</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>some college</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>master's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>associate's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>some college</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gender ethnicity  lunch  test_preparation_course  \\\n",
       "0  female         B      1                        0   \n",
       "1  female         C      1                        1   \n",
       "2  female         B      1                        0   \n",
       "3    male         A      0                        0   \n",
       "4    male         C      1                        0   \n",
       "\n",
       "  parental_level_of_education  \n",
       "0           bachelor's degree  \n",
       "1                some college  \n",
       "2             master's degree  \n",
       "3          associate's degree  \n",
       "4                some college  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "civil-pound",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    1\n",
       "1    1\n",
       "2    1\n",
       "3    0\n",
       "4    1\n",
       "Name: have_no_problem, dtype: int64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "young-ready",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.33)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "soviet-moment",
   "metadata": {},
   "source": [
    "# Логистическая регрессия"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "honey-display",
   "metadata": {},
   "outputs": [],
   "source": [
    "log_reg = glm(\n",
    "    f'have_no_problem ~ {\" + \".join(features) if len(features) else \"1\"}',\n",
    "    data=pd.concat([X_train, y_train], axis=1),\n",
    "    family=sm.families.Binomial()).fit()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "athletic-nicholas",
   "metadata": {},
   "source": [
    "# Сводная таблица"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "prescribed-steering",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Generalized Linear Model Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>    <td>have_no_problem</td> <th>  No. Observations:  </th>  <td>   670</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                  <td>GLM</td>       <th>  Df Residuals:      </th>  <td>   657</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model Family:</th>        <td>Binomial</td>     <th>  Df Model:          </th>  <td>    12</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Link Function:</th>         <td>logit</td>      <th>  Scale:             </th> <td>  1.0000</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                <td>IRLS</td>       <th>  Log-Likelihood:    </th> <td> -298.54</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Sat, 27 Mar 2021</td> <th>  Deviance:          </th> <td>  597.08</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>18:09:53</td>     <th>  Pearson chi2:      </th>  <td>  650.</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>No. Iterations:</th>          <td>5</td>        <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>                     </th>     <td> </td>   \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "                          <td></td>                            <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>                                        <td>   -0.0932</td> <td>    0.410</td> <td>   -0.227</td> <td> 0.820</td> <td>   -0.897</td> <td>    0.711</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>gender[T.male]</th>                                   <td>   -0.4913</td> <td>    0.208</td> <td>   -2.364</td> <td> 0.018</td> <td>   -0.899</td> <td>   -0.084</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ethnicity[T.B]</th>                                   <td>    0.3968</td> <td>    0.377</td> <td>    1.052</td> <td> 0.293</td> <td>   -0.342</td> <td>    1.136</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ethnicity[T.C]</th>                                   <td>    0.8296</td> <td>    0.362</td> <td>    2.291</td> <td> 0.022</td> <td>    0.120</td> <td>    1.539</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ethnicity[T.D]</th>                                   <td>    1.2594</td> <td>    0.388</td> <td>    3.250</td> <td> 0.001</td> <td>    0.500</td> <td>    2.019</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ethnicity[T.E]</th>                                   <td>    0.6987</td> <td>    0.419</td> <td>    1.667</td> <td> 0.095</td> <td>   -0.123</td> <td>    1.520</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>parental_level_of_education[T.bachelor's degree]</th> <td>    0.8331</td> <td>    0.469</td> <td>    1.777</td> <td> 0.076</td> <td>   -0.086</td> <td>    1.752</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>parental_level_of_education[T.high school]</th>       <td>   -0.4557</td> <td>    0.304</td> <td>   -1.500</td> <td> 0.134</td> <td>   -1.051</td> <td>    0.140</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>parental_level_of_education[T.master's degree]</th>   <td>    0.3363</td> <td>    0.542</td> <td>    0.620</td> <td> 0.535</td> <td>   -0.727</td> <td>    1.400</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>parental_level_of_education[T.some college]</th>      <td>   -0.1194</td> <td>    0.308</td> <td>   -0.387</td> <td> 0.699</td> <td>   -0.724</td> <td>    0.485</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>parental_level_of_education[T.some high school]</th>  <td>   -0.6871</td> <td>    0.316</td> <td>   -2.177</td> <td> 0.029</td> <td>   -1.306</td> <td>   -0.069</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>lunch</th>                                            <td>    1.3719</td> <td>    0.209</td> <td>    6.576</td> <td> 0.000</td> <td>    0.963</td> <td>    1.781</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>test_preparation_course</th>                          <td>    1.0527</td> <td>    0.240</td> <td>    4.395</td> <td> 0.000</td> <td>    0.583</td> <td>    1.522</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                 Generalized Linear Model Regression Results                  \n",
       "==============================================================================\n",
       "Dep. Variable:        have_no_problem   No. Observations:                  670\n",
       "Model:                            GLM   Df Residuals:                      657\n",
       "Model Family:                Binomial   Df Model:                           12\n",
       "Link Function:                  logit   Scale:                          1.0000\n",
       "Method:                          IRLS   Log-Likelihood:                -298.54\n",
       "Date:                Sat, 27 Mar 2021   Deviance:                       597.08\n",
       "Time:                        18:09:53   Pearson chi2:                     650.\n",
       "No. Iterations:                     5                                         \n",
       "Covariance Type:            nonrobust                                         \n",
       "====================================================================================================================\n",
       "                                                       coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------------------------------------\n",
       "Intercept                                           -0.0932      0.410     -0.227      0.820      -0.897       0.711\n",
       "gender[T.male]                                      -0.4913      0.208     -2.364      0.018      -0.899      -0.084\n",
       "ethnicity[T.B]                                       0.3968      0.377      1.052      0.293      -0.342       1.136\n",
       "ethnicity[T.C]                                       0.8296      0.362      2.291      0.022       0.120       1.539\n",
       "ethnicity[T.D]                                       1.2594      0.388      3.250      0.001       0.500       2.019\n",
       "ethnicity[T.E]                                       0.6987      0.419      1.667      0.095      -0.123       1.520\n",
       "parental_level_of_education[T.bachelor's degree]     0.8331      0.469      1.777      0.076      -0.086       1.752\n",
       "parental_level_of_education[T.high school]          -0.4557      0.304     -1.500      0.134      -1.051       0.140\n",
       "parental_level_of_education[T.master's degree]       0.3363      0.542      0.620      0.535      -0.727       1.400\n",
       "parental_level_of_education[T.some college]         -0.1194      0.308     -0.387      0.699      -0.724       0.485\n",
       "parental_level_of_education[T.some high school]     -0.6871      0.316     -2.177      0.029      -1.306      -0.069\n",
       "lunch                                                1.3719      0.209      6.576      0.000       0.963       1.781\n",
       "test_preparation_course                              1.0527      0.240      4.395      0.000       0.583       1.522\n",
       "====================================================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "log_reg.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "chief-islam",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = log_reg.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "smooth-introduction",
   "metadata": {},
   "source": [
    "# Выявленные студенты в группе риска"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "enhanced-hanging",
   "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>gender</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>lunch</th>\n",
       "      <th>test_preparation_course</th>\n",
       "      <th>parental_level_of_education</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>568</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>872</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>associate's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>981</th>\n",
       "      <td>male</td>\n",
       "      <td>D</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>male</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>some high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>425</th>\n",
       "      <td>female</td>\n",
       "      <td>C</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some college</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>218</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>723</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>422</th>\n",
       "      <td>female</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>master's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>892</th>\n",
       "      <td>female</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>master's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>705</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>bachelor's degree</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>330 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender ethnicity  lunch  test_preparation_course  \\\n",
       "568    male         B      0                        0   \n",
       "872    male         B      1                        1   \n",
       "981    male         D      1                        0   \n",
       "210    male         D      0                        1   \n",
       "425  female         C      0                        0   \n",
       "..      ...       ...    ...                      ...   \n",
       "218    male         B      0                        0   \n",
       "723    male         C      1                        0   \n",
       "422  female         D      0                        1   \n",
       "892  female         A      0                        0   \n",
       "705    male         A      0                        0   \n",
       "\n",
       "    parental_level_of_education  \n",
       "568                 high school  \n",
       "872          associate's degree  \n",
       "981            some high school  \n",
       "210            some high school  \n",
       "425                some college  \n",
       "..                          ...  \n",
       "218                 high school  \n",
       "723                 high school  \n",
       "422             master's degree  \n",
       "892             master's degree  \n",
       "705           bachelor's degree  \n",
       "\n",
       "[330 rows x 5 columns]"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "political-relative",
   "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>gender</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>lunch</th>\n",
       "      <th>test_preparation_course</th>\n",
       "      <th>parental_level_of_education</th>\n",
       "      <th>predictions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>568</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.344463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some college</td>\n",
       "      <td>0.423813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>master's degree</td>\n",
       "      <td>0.438261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>830</th>\n",
       "      <td>female</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some college</td>\n",
       "      <td>0.447041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>associate's degree</td>\n",
       "      <td>0.357896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>male</td>\n",
       "      <td>D</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.496948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>444</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.218988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>228</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.218988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>731</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.218988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>882</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.462025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>490</th>\n",
       "      <td>female</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>associate's degree</td>\n",
       "      <td>0.476710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>427</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.391268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>952</th>\n",
       "      <td>female</td>\n",
       "      <td>E</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.479617</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>395</th>\n",
       "      <td>male</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.261103</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>396</th>\n",
       "      <td>female</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.462025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.447530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>442</th>\n",
       "      <td>female</td>\n",
       "      <td>A</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.314256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>887</th>\n",
       "      <td>male</td>\n",
       "      <td>C</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.447530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>some high school</td>\n",
       "      <td>0.294257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>218</th>\n",
       "      <td>male</td>\n",
       "      <td>B</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>high school</td>\n",
       "      <td>0.344463</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gender ethnicity  lunch  test_preparation_course  \\\n",
       "568    male         B      0                        0   \n",
       "157    male         B      0                        0   \n",
       "25     male         A      0                        0   \n",
       "830  female         A      0                        0   \n",
       "3      male         A      0                        0   \n",
       "301    male         D      0                        0   \n",
       "444    male         A      0                        0   \n",
       "228    male         A      0                        0   \n",
       "731    male         A      0                        0   \n",
       "882  female         B      0                        0   \n",
       "490  female         A      0                        0   \n",
       "427    male         C      0                        0   \n",
       "952  female         E      0                        0   \n",
       "395    male         A      0                        0   \n",
       "396  female         B      0                        0   \n",
       "91     male         C      0                        0   \n",
       "442  female         A      0                        0   \n",
       "887    male         C      0                        0   \n",
       "188    male         B      0                        0   \n",
       "218    male         B      0                        0   \n",
       "\n",
       "    parental_level_of_education  predictions  \n",
       "568                 high school     0.344463  \n",
       "157                some college     0.423813  \n",
       "25              master's degree     0.438261  \n",
       "830                some college     0.447041  \n",
       "3            associate's degree     0.357896  \n",
       "301            some high school     0.496948  \n",
       "444            some high school     0.218988  \n",
       "228            some high school     0.218988  \n",
       "731            some high school     0.218988  \n",
       "882                 high school     0.462025  \n",
       "490          associate's degree     0.476710  \n",
       "427            some high school     0.391268  \n",
       "952            some high school     0.479617  \n",
       "395                 high school     0.261103  \n",
       "396                 high school     0.462025  \n",
       "91                  high school     0.447530  \n",
       "442            some high school     0.314256  \n",
       "887                 high school     0.447530  \n",
       "188            some high school     0.294257  \n",
       "218                 high school     0.344463  "
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "students_with_problems = X_test.copy()\n",
    "students_with_problems['predictions'] = predictions\n",
    "students_with_problems[students_with_problems.iloc[:,-1] < .5]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bridal-manitoba",
   "metadata": {},
   "source": [
    "# Метрики качества модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "subsequent-daily",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.35      0.11      0.17        64\n",
      "           1       0.82      0.95      0.88       266\n",
      "\n",
      "    accuracy                           0.79       330\n",
      "   macro avg       0.58      0.53      0.52       330\n",
      "weighted avg       0.73      0.79      0.74       330\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, round(predictions)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "endless-breeding",
   "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": [
    "print_roc(y_test, predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "therapeutic-cooler",
   "metadata": {},
   "source": [
    "# Дерево решений"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "average-indication",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from sklearn import tree\n",
    "\n",
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "artistic-italian",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.get_dummies(X)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.33)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "great-brunei",
   "metadata": {},
   "outputs": [],
   "source": [
    "#dt = DecisionTreeClassifier()\n",
    "# parameters = {\n",
    "#     'criterion': ['entropy', 'gini'],\n",
    "#     'max_depth': range(1, 10),\n",
    "#     'min_samples_split': range(2, 50),\n",
    "#     'min_samples_leaf': range(1, 50),\n",
    "# }\n",
    "\n",
    "# search = GridSearchCV(dt, parameters,\n",
    "#                       cv=5,\n",
    "#                       n_jobs=-1).fit(X_train, y_train)\n",
    "# search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "expanded-title",
   "metadata": {},
   "outputs": [],
   "source": [
    "#best_tree = search.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "boxed-pakistan",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_tree = DecisionTreeClassifier(\n",
    "    criterion='entropy',\n",
    "    max_depth=3,\n",
    "    min_samples_leaf=1,\n",
    "    min_samples_split=2\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "abandoned-luxembourg",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(criterion='entropy', max_depth=3)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_tree.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "nuclear-station",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<svg xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\" width=\"1369pt\" height=\"433pt\" viewBox=\"0.00 0.00 1369.00 433.00\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 429)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"white\" stroke=\"transparent\" points=\"-4,4 -4,-429 1365,-429 1365,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\">\n",
       "<title>0</title>\n",
       "<polygon fill=\"#6bb6ec\" stroke=\"black\" points=\"746,-425 589,-425 589,-342 746,-342 746,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"667.5\" y=\"-409.8\" font-family=\"Times,serif\" font-size=\"14.00\">lunch &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"667.5\" y=\"-394.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.728</text>\n",
       "<text text-anchor=\"middle\" x=\"667.5\" y=\"-379.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 670</text>\n",
       "<text text-anchor=\"middle\" x=\"667.5\" y=\"-364.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [136, 534]</text>\n",
       "<text text-anchor=\"middle\" x=\"667.5\" y=\"-349.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\">\n",
       "<title>1</title>\n",
       "<polygon fill=\"#98ccf2\" stroke=\"black\" points=\"659,-306 414,-306 414,-223 659,-223 659,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"536.5\" y=\"-290.8\" font-family=\"Times,serif\" font-size=\"14.00\">test_preparation_course &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"536.5\" y=\"-275.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.91</text>\n",
       "<text text-anchor=\"middle\" x=\"536.5\" y=\"-260.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 237</text>\n",
       "<text text-anchor=\"middle\" x=\"536.5\" y=\"-245.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [77, 160]</text>\n",
       "<text text-anchor=\"middle\" x=\"536.5\" y=\"-230.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\">\n",
       "<title>0-&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M622.05,-341.91C611.58,-332.56 600.36,-322.54 589.6,-312.93\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"591.66,-310.07 581.87,-306.02 587,-315.29 591.66,-310.07\"/>\n",
       "<text text-anchor=\"middle\" x=\"583.28\" y=\"-327.28\" font-family=\"Times,serif\" font-size=\"14.00\">True</text>\n",
       "</g>\n",
       "<!-- 8 -->\n",
       "<g id=\"node9\" class=\"node\">\n",
       "<title>8</title>\n",
       "<polygon fill=\"#58ace9\" stroke=\"black\" points=\"922,-306 677,-306 677,-223 922,-223 922,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"799.5\" y=\"-290.8\" font-family=\"Times,serif\" font-size=\"14.00\">test_preparation_course &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"799.5\" y=\"-275.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.574</text>\n",
       "<text text-anchor=\"middle\" x=\"799.5\" y=\"-260.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 433</text>\n",
       "<text text-anchor=\"middle\" x=\"799.5\" y=\"-245.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [59, 374]</text>\n",
       "<text text-anchor=\"middle\" x=\"799.5\" y=\"-230.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;8 -->\n",
       "<g id=\"edge8\" class=\"edge\">\n",
       "<title>0-&gt;8</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M713.3,-341.91C723.84,-332.56 735.15,-322.54 745.99,-312.93\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"748.62,-315.27 753.78,-306.02 743.98,-310.03 748.62,-315.27\"/>\n",
       "<text text-anchor=\"middle\" x=\"752.28\" y=\"-327.28\" font-family=\"Times,serif\" font-size=\"14.00\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\">\n",
       "<title>2</title>\n",
       "<polygon fill=\"#b2d9f5\" stroke=\"black\" points=\"441,-187 42,-187 42,-104 441,-104 441,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"241.5\" y=\"-171.8\" font-family=\"Times,serif\" font-size=\"14.00\">parental_level_of_education_some high school &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"241.5\" y=\"-156.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.957</text>\n",
       "<text text-anchor=\"middle\" x=\"241.5\" y=\"-141.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 148</text>\n",
       "<text text-anchor=\"middle\" x=\"241.5\" y=\"-126.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [56, 92]</text>\n",
       "<text text-anchor=\"middle\" x=\"241.5\" y=\"-111.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\">\n",
       "<title>1-&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M434.15,-222.91C407.95,-212.52 379.65,-201.29 353.01,-190.73\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"354.25,-187.45 343.67,-187.02 351.67,-193.96 354.25,-187.45\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\">\n",
       "<title>5</title>\n",
       "<polygon fill=\"#76bbed\" stroke=\"black\" points=\"616,-187 459,-187 459,-104 616,-104 616,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"537.5\" y=\"-171.8\" font-family=\"Times,serif\" font-size=\"14.00\">ethnicity_A &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"537.5\" y=\"-156.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.788</text>\n",
       "<text text-anchor=\"middle\" x=\"537.5\" y=\"-141.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 89</text>\n",
       "<text text-anchor=\"middle\" x=\"537.5\" y=\"-126.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [21, 68]</text>\n",
       "<text text-anchor=\"middle\" x=\"537.5\" y=\"-111.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\">\n",
       "<title>1-&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M536.85,-222.91C536.92,-214.65 536.99,-205.86 537.07,-197.3\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"540.57,-197.05 537.15,-187.02 533.57,-196.99 540.57,-197.05\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\">\n",
       "<title>3</title>\n",
       "<polygon fill=\"#a5d2f3\" stroke=\"black\" points=\"157,-68 0,-68 0,0 157,0 157,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"78.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.936</text>\n",
       "<text text-anchor=\"middle\" x=\"78.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 125</text>\n",
       "<text text-anchor=\"middle\" x=\"78.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [44, 81]</text>\n",
       "<text text-anchor=\"middle\" x=\"78.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\">\n",
       "<title>2-&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M180.8,-103.73C166.2,-93.92 150.64,-83.46 136.18,-73.75\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"138.06,-70.8 127.81,-68.13 134.16,-76.61 138.06,-70.8\"/>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\">\n",
       "<title>4</title>\n",
       "<polygon fill=\"#fdf5ef\" stroke=\"black\" points=\"309.5,-68 175.5,-68 175.5,0 309.5,0 309.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"242.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.999</text>\n",
       "<text text-anchor=\"middle\" x=\"242.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 23</text>\n",
       "<text text-anchor=\"middle\" x=\"242.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [12, 11]</text>\n",
       "<text text-anchor=\"middle\" x=\"242.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = Problem</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\">\n",
       "<title>2-&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M241.87,-103.73C241.95,-95.52 242.03,-86.86 242.1,-78.56\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"245.6,-78.33 242.2,-68.3 238.6,-78.27 245.6,-78.33\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\">\n",
       "<title>6</title>\n",
       "<polygon fill=\"#6db7ec\" stroke=\"black\" points=\"498,-68 341,-68 341,0 498,0 498,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"419.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.736</text>\n",
       "<text text-anchor=\"middle\" x=\"419.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 82</text>\n",
       "<text text-anchor=\"middle\" x=\"419.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [17, 65]</text>\n",
       "<text text-anchor=\"middle\" x=\"419.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\">\n",
       "<title>5-&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M493.56,-103.73C483.54,-94.42 472.89,-84.54 462.89,-75.26\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"465.1,-72.54 455.38,-68.3 460.33,-77.67 465.1,-72.54\"/>\n",
       "</g>\n",
       "<!-- 7 -->\n",
       "<g id=\"node8\" class=\"node\">\n",
       "<title>7</title>\n",
       "<polygon fill=\"#f8e0ce\" stroke=\"black\" points=\"650.5,-68 516.5,-68 516.5,0 650.5,0 650.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"583.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.985</text>\n",
       "<text text-anchor=\"middle\" x=\"583.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 7</text>\n",
       "<text text-anchor=\"middle\" x=\"583.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [4, 3]</text>\n",
       "<text text-anchor=\"middle\" x=\"583.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = Problem</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;7 -->\n",
       "<g id=\"edge7\" class=\"edge\">\n",
       "<title>5-&gt;7</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M554.63,-103.73C558.19,-95.24 561.96,-86.28 565.55,-77.73\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"568.86,-78.87 569.51,-68.3 562.41,-76.16 568.86,-78.87\"/>\n",
       "</g>\n",
       "<!-- 9 -->\n",
       "<g id=\"node10\" class=\"node\">\n",
       "<title>9</title>\n",
       "<polygon fill=\"#65b3eb\" stroke=\"black\" points=\"886,-187 709,-187 709,-104 886,-104 886,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"797.5\" y=\"-171.8\" font-family=\"Times,serif\" font-size=\"14.00\">gender_female &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"797.5\" y=\"-156.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.685</text>\n",
       "<text text-anchor=\"middle\" x=\"797.5\" y=\"-141.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 285</text>\n",
       "<text text-anchor=\"middle\" x=\"797.5\" y=\"-126.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [52, 233]</text>\n",
       "<text text-anchor=\"middle\" x=\"797.5\" y=\"-111.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;9 -->\n",
       "<g id=\"edge9\" class=\"edge\">\n",
       "<title>8-&gt;9</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M798.81,-222.91C798.66,-214.56 798.51,-205.67 798.36,-197.02\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"801.86,-196.96 798.19,-187.02 794.86,-197.08 801.86,-196.96\"/>\n",
       "</g>\n",
       "<!-- 12 -->\n",
       "<g id=\"node13\" class=\"node\">\n",
       "<title>12</title>\n",
       "<polygon fill=\"#43a2e6\" stroke=\"black\" points=\"1311,-187 904,-187 904,-104 1311,-104 1311,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-171.8\" font-family=\"Times,serif\" font-size=\"14.00\">parental_level_of_education_associate's degree &lt;= 0.5</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-156.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.275</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-141.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 148</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-126.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [7, 141]</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-111.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;12 -->\n",
       "<g id=\"edge12\" class=\"edge\">\n",
       "<title>8-&gt;12</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M906.36,-222.91C933.83,-212.47 963.52,-201.19 991.44,-190.59\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"992.72,-193.84 1000.83,-187.02 990.24,-187.3 992.72,-193.84\"/>\n",
       "</g>\n",
       "<!-- 10 -->\n",
       "<g id=\"node11\" class=\"node\">\n",
       "<title>10</title>\n",
       "<polygon fill=\"#78bced\" stroke=\"black\" points=\"833,-68 676,-68 676,0 833,0 833,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"754.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.796</text>\n",
       "<text text-anchor=\"middle\" x=\"754.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 137</text>\n",
       "<text text-anchor=\"middle\" x=\"754.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [33, 104]</text>\n",
       "<text text-anchor=\"middle\" x=\"754.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 9&#45;&gt;10 -->\n",
       "<g id=\"edge10\" class=\"edge\">\n",
       "<title>9-&gt;10</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M781.49,-103.73C778.16,-95.24 774.64,-86.28 771.28,-77.73\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"774.49,-76.33 767.58,-68.3 767.97,-78.89 774.49,-76.33\"/>\n",
       "</g>\n",
       "<!-- 11 -->\n",
       "<g id=\"node12\" class=\"node\">\n",
       "<title>11</title>\n",
       "<polygon fill=\"#56abe9\" stroke=\"black\" points=\"1008,-68 851,-68 851,0 1008,0 1008,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"929.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.553</text>\n",
       "<text text-anchor=\"middle\" x=\"929.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 148</text>\n",
       "<text text-anchor=\"middle\" x=\"929.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [19, 129]</text>\n",
       "<text text-anchor=\"middle\" x=\"929.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 9&#45;&gt;11 -->\n",
       "<g id=\"edge11\" class=\"edge\">\n",
       "<title>9-&gt;11</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M846.65,-103.73C858.09,-94.24 870.24,-84.16 881.62,-74.72\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"883.9,-77.38 889.36,-68.3 879.43,-71.99 883.9,-77.38\"/>\n",
       "</g>\n",
       "<!-- 13 -->\n",
       "<g id=\"node14\" class=\"node\">\n",
       "<title>13</title>\n",
       "<polygon fill=\"#46a4e7\" stroke=\"black\" points=\"1186,-68 1029,-68 1029,0 1186,0 1186,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.337</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 112</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [7, 105]</text>\n",
       "<text text-anchor=\"middle\" x=\"1107.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;13 -->\n",
       "<g id=\"edge13\" class=\"edge\">\n",
       "<title>12-&gt;13</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M1107.5,-103.73C1107.5,-95.52 1107.5,-86.86 1107.5,-78.56\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"1111,-78.3 1107.5,-68.3 1104,-78.3 1111,-78.3\"/>\n",
       "</g>\n",
       "<!-- 14 -->\n",
       "<g id=\"node15\" class=\"node\">\n",
       "<title>14</title>\n",
       "<polygon fill=\"#399de5\" stroke=\"black\" points=\"1361,-68 1204,-68 1204,0 1361,0 1361,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"1282.5\" y=\"-52.8\" font-family=\"Times,serif\" font-size=\"14.00\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"1282.5\" y=\"-37.8\" font-family=\"Times,serif\" font-size=\"14.00\">samples = 36</text>\n",
       "<text text-anchor=\"middle\" x=\"1282.5\" y=\"-22.8\" font-family=\"Times,serif\" font-size=\"14.00\">value = [0, 36]</text>\n",
       "<text text-anchor=\"middle\" x=\"1282.5\" y=\"-7.8\" font-family=\"Times,serif\" font-size=\"14.00\">class = No problem</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;14 -->\n",
       "<g id=\"edge14\" class=\"edge\">\n",
       "<title>12-&gt;14</title>\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M1172.66,-103.73C1188.49,-93.82 1205.36,-83.27 1221.01,-73.48\"/>\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"1222.94,-76.4 1229.56,-68.13 1219.23,-70.46 1222.94,-76.4\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>"
      ],
      "text/plain": [
       "<IPython.core.display.SVG object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "graph = Source(tree.export_graphviz(\n",
    "    best_tree,\n",
    "    out_file=None,\n",
    "    feature_names=list(X),\n",
    "    class_names=['Problem', 'No problem'],\n",
    "    filled=True\n",
    "))\n",
    "\n",
    "display(SVG(graph.pipe(format='svg')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "controversial-upset",
   "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>feature_importances</th>\n",
       "      <th>features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.452332</td>\n",
       "      <td>lunch</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.318884</td>\n",
       "      <td>test_preparation_course</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.084810</td>\n",
       "      <td>gender_female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.055979</td>\n",
       "      <td>parental_level_of_education_associate's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.055682</td>\n",
       "      <td>ethnicity_A</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.032312</td>\n",
       "      <td>parental_level_of_education_some high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>gender_male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>ethnicity_B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>ethnicity_C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>ethnicity_D</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>ethnicity_E</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>parental_level_of_education_bachelor's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>parental_level_of_education_high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>parental_level_of_education_master's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>parental_level_of_education_some college</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    feature_importances                                        features\n",
       "0              0.452332                                           lunch\n",
       "1              0.318884                         test_preparation_course\n",
       "2              0.084810                                   gender_female\n",
       "9              0.055979  parental_level_of_education_associate's degree\n",
       "4              0.055682                                     ethnicity_A\n",
       "14             0.032312    parental_level_of_education_some high school\n",
       "3              0.000000                                     gender_male\n",
       "5              0.000000                                     ethnicity_B\n",
       "6              0.000000                                     ethnicity_C\n",
       "7              0.000000                                     ethnicity_D\n",
       "8              0.000000                                     ethnicity_E\n",
       "10             0.000000   parental_level_of_education_bachelor's degree\n",
       "11             0.000000         parental_level_of_education_high school\n",
       "12             0.000000     parental_level_of_education_master's degree\n",
       "13             0.000000        parental_level_of_education_some college"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_importances = best_tree.feature_importances_\n",
    "feature_importances_df = pd.DataFrame({\n",
    "    'feature_importances': feature_importances,\n",
    "    'features': list(X_train)\n",
    "})\n",
    "feature_importances_df.sort_values('feature_importances', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "aggregate-accessory",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = best_tree.predict_proba(X_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "fossil-boating",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.56      0.14      0.22        72\n",
      "           1       0.80      0.97      0.88       258\n",
      "\n",
      "    accuracy                           0.79       330\n",
      "   macro avg       0.68      0.55      0.55       330\n",
      "weighted avg       0.75      0.79      0.73       330\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, predictions.round()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "fleet-brisbane",
   "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": [
    "print_roc(y_test, predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ordinary-dubai",
   "metadata": {},
   "source": [
    "# Случайный лес"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "together-college",
   "metadata": {},
   "outputs": [],
   "source": [
    "# rf = RandomForestClassifier()\n",
    "# parameters = {\n",
    "#     'max_depth': range(1, 5),\n",
    "#     'min_samples_split': range(2, 40, 10),\n",
    "#     'min_samples_leaf': range(2, 40, 10),\n",
    "#     'n_estimators': range(30, 40, 2),\n",
    "# }\n",
    "\n",
    "# search = GridSearchCV(rf, parameters,\n",
    "#                       cv=5,\n",
    "#                       n_jobs=-1).fit(X_train, y_train)\n",
    "# search.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "joint-scott",
   "metadata": {},
   "outputs": [],
   "source": [
    "# best_forest = search.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "cellular-breach",
   "metadata": {},
   "outputs": [],
   "source": [
    "best_forest = RandomForestClassifier(\n",
    "    max_depth=4,\n",
    "    min_samples_leaf=2, \n",
    "    min_samples_split=22,\n",
    "    n_estimators=36\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "deluxe-cookie",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=4, min_samples_leaf=2, min_samples_split=22,\n",
       "                       n_estimators=36)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_forest.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "opened-investigation",
   "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>feature_importances</th>\n",
       "      <th>features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.314601</td>\n",
       "      <td>lunch</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.182946</td>\n",
       "      <td>test_preparation_course</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.089060</td>\n",
       "      <td>gender_male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.060349</td>\n",
       "      <td>gender_female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.049729</td>\n",
       "      <td>parental_level_of_education_high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.042057</td>\n",
       "      <td>parental_level_of_education_some high school</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.039412</td>\n",
       "      <td>parental_level_of_education_master's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.034065</td>\n",
       "      <td>ethnicity_E</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.030350</td>\n",
       "      <td>parental_level_of_education_some college</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.028723</td>\n",
       "      <td>ethnicity_B</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.028643</td>\n",
       "      <td>ethnicity_A</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.027487</td>\n",
       "      <td>ethnicity_D</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.025029</td>\n",
       "      <td>parental_level_of_education_bachelor's degree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.024926</td>\n",
       "      <td>ethnicity_C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.022623</td>\n",
       "      <td>parental_level_of_education_associate's degree</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    feature_importances                                        features\n",
       "0              0.314601                                           lunch\n",
       "1              0.182946                         test_preparation_course\n",
       "3              0.089060                                     gender_male\n",
       "2              0.060349                                   gender_female\n",
       "11             0.049729         parental_level_of_education_high school\n",
       "14             0.042057    parental_level_of_education_some high school\n",
       "12             0.039412     parental_level_of_education_master's degree\n",
       "8              0.034065                                     ethnicity_E\n",
       "13             0.030350        parental_level_of_education_some college\n",
       "5              0.028723                                     ethnicity_B\n",
       "4              0.028643                                     ethnicity_A\n",
       "7              0.027487                                     ethnicity_D\n",
       "10             0.025029   parental_level_of_education_bachelor's degree\n",
       "6              0.024926                                     ethnicity_C\n",
       "9              0.022623  parental_level_of_education_associate's degree"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_importances = best_forest.feature_importances_\n",
    "feature_importances_df = pd.DataFrame({\n",
    "    'feature_importances': feature_importances,\n",
    "    'features': list(X_train)\n",
    "})\n",
    "feature_importances_df.sort_values('feature_importances', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "connected-methodology",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = best_forest.predict_proba(X_test)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "private-messaging",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00        72\n",
      "           1       0.78      1.00      0.88       258\n",
      "\n",
      "    accuracy                           0.78       330\n",
      "   macro avg       0.39      0.50      0.44       330\n",
      "weighted avg       0.61      0.78      0.69       330\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/starlightx/.local/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/starlightx/.local/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/starlightx/.local/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1245: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, predictions.round()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "numerical-slovenia",
   "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": [
    "print_roc(y_test, predictions)"
   ]
  }
 ],
 "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": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}