enem_notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ENEM - Success or Unsuccess\n",
    "### An Classification problem\n",
    "\n",
    "Starting slowly, we intend to define whether a participant would be successful or not on ENEM tests based on their socioeconomical status, so, we chose to use the learning model `LinearSVG`, a Linear Support Vector Classification to predict this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "VKs_qGdJkQeR"
   },
   "outputs": [],
   "source": [
    "#Import libraries\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.pipeline import make_pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "peHkdLnqkQee"
   },
   "outputs": [],
   "source": [
    "#Import Filtered Dataset containing the selected features\n",
    "dataEnem = pd.read_csv (\"DataEnemQuestionsParsed.csv\", encoding='ISO-8859-1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 241
    },
    "executionInfo": {
     "elapsed": 1591,
     "status": "ok",
     "timestamp": 1601915294931,
     "user": {
      "displayName": "OZIEL JUNIOR",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Ggz0UUkl87dG84DhMt9Y5LE_VTSr34DlL8zLb9G=s64",
      "userId": "07062214448555350795"
     },
     "user_tz": 180
    },
    "id": "dxk6abvTkQem",
    "outputId": "51273fec-b42e-48cb-9281-81a376ead7a1"
   },
   "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>Unnamed: 0</th>\n",
       "      <th>Unnamed: 0.1</th>\n",
       "      <th>NU_INSCRICAO</th>\n",
       "      <th>TP_NACIONALIDADE</th>\n",
       "      <th>TP_ST_CONCLUSAO</th>\n",
       "      <th>TP_ESCOLA</th>\n",
       "      <th>TP_ENSINO</th>\n",
       "      <th>SG_UF_ESC</th>\n",
       "      <th>TP_DEPENDENCIA_ADM_ESC</th>\n",
       "      <th>TP_LOCALIZACAO_ESC</th>\n",
       "      <th>...</th>\n",
       "      <th>Q016</th>\n",
       "      <th>Q017</th>\n",
       "      <th>Q018</th>\n",
       "      <th>Q019</th>\n",
       "      <th>Q020</th>\n",
       "      <th>Q021</th>\n",
       "      <th>Q022</th>\n",
       "      <th>Q023</th>\n",
       "      <th>Q024</th>\n",
       "      <th>Q025</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>190001004629</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>190001004634</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>15</td>\n",
       "      <td>190001004642</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>PA</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>190001004643</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>24</td>\n",
       "      <td>190001004651</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>PA</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 45 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  Unnamed: 0.1  NU_INSCRICAO  TP_NACIONALIDADE  TP_ST_CONCLUSAO  \\\n",
       "0           0             2  190001004629                 1                2   \n",
       "1           1             7  190001004634                 1                2   \n",
       "2           2            15  190001004642                 1                2   \n",
       "3           3            16  190001004643                 1                2   \n",
       "4           4            24  190001004651                 1                2   \n",
       "\n",
       "   TP_ESCOLA  TP_ENSINO SG_UF_ESC  TP_DEPENDENCIA_ADM_ESC  TP_LOCALIZACAO_ESC  \\\n",
       "0          3        1.0        PA                     4.0                 1.0   \n",
       "1          2        1.0        PA                     2.0                 1.0   \n",
       "2          2        NaN        PA                     2.0                 1.0   \n",
       "3          2        1.0        PA                     2.0                 1.0   \n",
       "4          2        1.0        PA                     2.0                 1.0   \n",
       "\n",
       "   ...  Q016  Q017  Q018  Q019  Q020  Q021  Q022  Q023  Q024  Q025  \n",
       "0  ...     2     1     1     4     2     2     4     1     3     2  \n",
       "1  ...     1     1     1     2     1     1     2     1     1     1  \n",
       "2  ...     1     1     1     2     2     1     4     1     1     1  \n",
       "3  ...     1     1     1     4     1     1     3     1     1     1  \n",
       "4  ...     2     1     1     2     1     1     4     1     1     2  \n",
       "\n",
       "[5 rows x 45 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "brDataEnem = dataEnem\n",
    "brDataEnem.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Mapping questions inputs\n",
    "We chose to map the inputs using the alphabetical order so that there is no hierarchical loss in the value of the results according to the data dictionary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 102
    },
    "id": "M72gH8yZnl96",
    "outputId": "315bfdfc-8a63-4f89-a105-2a2bf32e6ea2"
   },
   "outputs": [],
   "source": [
    "#Parsing Questions to Numbers\n",
    "#Q005 already has numbers as input\n",
    "#questions = ['Q001', 'Q002', 'Q003', 'Q004', 'Q006', 'Q007', 'Q008', \n",
    "     #        'Q009', 'Q010', 'Q011', 'Q012', 'Q013', 'Q014','Q015', 'Q016', \n",
    "   #          'Q017', 'Q018', 'Q019', 'Q020', 'Q021', 'Q022','Q023', 'Q024', 'Q025']\n",
    "\n",
    "#for i in questions:\n",
    "#    brDataEnem[i] = [ ord(x) - 64 for x in brDataEnem[i] ]\n",
    "#brDataEnem.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Success and Unsuccess definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "_B4dQe9FkQfl"
   },
   "outputs": [],
   "source": [
    "#Finds the median for every test\n",
    "medianRe = brDataEnem['NU_NOTA_REDACAO'].median()\n",
    "medianCN = brDataEnem['NU_NOTA_CN'].median()\n",
    "medianCH = brDataEnem['NU_NOTA_CH'].median()\n",
    "medianLC = brDataEnem['NU_NOTA_LC'].median()\n",
    "medianMT = brDataEnem['NU_NOTA_MT'].median()\n",
    "   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating an aditional boolean column for each test that defines it's success"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "id": "Oa3D5xabkQft"
   },
   "outputs": [],
   "source": [
    "#Create a new column called brDataEnem.[TestName] where the value\n",
    "#is yes if the score is greater than the median and 0 otherwise.\n",
    "#1 = greater than median, 0 = lesser than median\n",
    "brDataEnem['Re'] = np.where(brDataEnem['NU_NOTA_REDACAO'] >= medianRe, 1, 0)\n",
    "brDataEnem['CN'] = np.where(brDataEnem['NU_NOTA_CN'] >= medianCN, 1, 0)\n",
    "brDataEnem['CH'] = np.where(brDataEnem['NU_NOTA_CH'] >= medianCH, 1, 0)\n",
    "brDataEnem['LC'] = np.where(brDataEnem['NU_NOTA_LC'] >= medianLC, 1, 0)\n",
    "brDataEnem['MT'] = np.where(brDataEnem['NU_NOTA_MT'] >= medianMT, 1, 0)\n",
    "\n",
    "#Create a series for baseline algorithm\n",
    "brDataEnem['BASELINE_TP_ESCOLA'] = np.where(brDataEnem['TP_ESCOLA'] == 2, 0, 1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "QIYub5ZfkQf1"
   },
   "outputs": [],
   "source": [
    "#Getting features and labels that will be used in classification\n",
    "X = brDataEnem[['TP_ESCOLA','Q001', 'Q002', 'Q003', 'Q004', 'Q005', 'Q006',\n",
    "                'Q007', 'Q008', 'Q009', 'Q010', 'Q011', 'Q012', 'Q013', 'Q014',\n",
    "                'Q015', 'Q016', 'Q017', 'Q018', 'Q019', 'Q020', 'Q021', 'Q022',\n",
    "                'Q023', 'Q024', 'Q025', 'TP_DEPENDENCIA_ADM_ESC']]\n",
    "y = brDataEnem['MT']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "executionInfo": {
     "elapsed": 46971,
     "status": "ok",
     "timestamp": 1601915396051,
     "user": {
      "displayName": "OZIEL JUNIOR",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Ggz0UUkl87dG84DhMt9Y5LE_VTSr34DlL8zLb9G=s64",
      "userId": "07062214448555350795"
     },
     "user_tz": 180
    },
    "id": "F3yB1jBEkQf8",
    "outputId": "5bf79483-7b85-4f62-ebf0-c916d862329d"
   },
   "outputs": [],
   "source": [
    "#Defining the training set \n",
    "SEED = 30\n",
    "treino_x, teste_x, treino_y, teste_y = train_test_split(X, y, random_state = SEED, test_size = 0.25)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    345752\n",
       "1    345666\n",
       "Name: MT, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Traning\n",
    "modelo = make_pipeline(StandardScaler(),LinearSVC(dual=False, max_iter=5000))\n",
    "modelo.fit(treino_x, treino_y)\n",
    "previsoes = modelo.predict(teste_x)\n",
    "acuracia = accuracy_score(teste_y, previsoes) * 100\n",
    "\n",
    "treino_y.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparting to baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 66.87%\n",
      "The baseline accuracy using the type of school was 49.95%\n",
      "The random baseline accuracy was 50.02%\n"
     ]
    }
   ],
   "source": [
    "def baseline_testing(teste_y):\n",
    "    #Baseline using the type of school\n",
    "    base_escola = brDataEnem['BASELINE_TP_ESCOLA'].to_numpy()\n",
    "    base_escola_teste = base_escola[:len(teste_y.index)]\n",
    "    acuracia_base = accuracy_score(teste_y, base_escola_teste) * 100\n",
    "    print(\"The baseline accuracy using the type of school was %.2f%%\" % acuracia_base)\n",
    "\n",
    "    #Baseline all ones\n",
    "    all_ones = np.ones((len(teste_y.index)))\n",
    "    acuracia_base = accuracy_score(teste_y, all_ones) * 100\n",
    "    print(\"The random baseline accuracy was %.2f%%\" % acuracia_base)\n",
    "\n",
    "print(\"Accuracy: %.2f%%\" % acuracia)\n",
    "baseline_testing(teste_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Features Relevance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "27\n",
      "['q001', 'q002', 'q003', 'q004', 'q005', 'q006', 'q007', 'q008', 'q009', 'q010', 'q011', 'q012', 'q013', 'q014', 'q015', 'q016', 'q017', 'q018', 'q019', 'q020', 'q021', 'q022', 'q023', 'q024', 'q025', 'tp_dependencia_adm_esc', 'tp_escola']\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "def plot_coefficients(classifier, feature_names, top_features=20):\n",
    " coef = classifier.coef_.ravel()\n",
    " top_positive_coefficients = np.argsort(coef)[-top_features:]\n",
    " top_negative_coefficients = np.argsort(coef)[:top_features]\n",
    " top_coefficients = np.hstack([top_negative_coefficients, top_positive_coefficients])\n",
    " \n",
    "# create plot\n",
    " plt.figure(figsize=(15, 5))\n",
    " colors = ['red' if c < 0 else 'blue' for c in coef[top_coefficients]]\n",
    " plt.bar(np.arange(2 * top_features), coef[top_coefficients], color=colors)\n",
    " feature_names = np.array(feature_names)\n",
    " plt.xticks(np.arange(1, 1 + 2 * top_features), feature_names[top_coefficients], rotation=60, ha='right')\n",
    " plt.show()\n",
    "\n",
    "cv = CountVectorizer()\n",
    "cv.fit(X)\n",
    "print(len(cv.vocabulary_))\n",
    "print(cv.get_feature_names())\n",
    "#X_train = cv.transform(data)\n",
    "\n",
    "svm = LinearSVC(dual=False, max_iter=5000)\n",
    "svm.fit(treino_x, treino_y)\n",
    "plot_coefficients(svm, cv.get_feature_names())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Languages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    345888\n",
       "0    345530\n",
       "Name: LC, dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = brDataEnem['LC']\n",
    "\n",
    "#Defining the training set \n",
    "SEED = 30\n",
    "treino_x, teste_x, treino_y, teste_y = train_test_split(X, y, random_state = SEED, test_size = 0.25)\n",
    "\n",
    "#Traning\n",
    "modelo = make_pipeline(StandardScaler(),LinearSVC(dual=False, max_iter=5000))\n",
    "modelo.fit(treino_x, treino_y)\n",
    "previsoes = modelo.predict(teste_x)\n",
    "acuracia = accuracy_score(teste_y, previsoes) * 100\n",
    "treino_y.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparing to baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 68.38%\n",
      "The baseline accuracy using the type of school was 49.93%\n",
      "The random baseline accuracy was 50.02%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\" % acuracia)\n",
    "baseline_testing(teste_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Natural Sciences\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    345826\n",
       "1    345592\n",
       "Name: CN, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = brDataEnem['CN']\n",
    "\n",
    "#Defining the training set \n",
    "SEED = 30\n",
    "treino_x, teste_x, treino_y, teste_y = train_test_split(X, y, random_state = SEED, test_size = 0.25)\n",
    "\n",
    "#Traning\n",
    "modelo = make_pipeline(StandardScaler(),LinearSVC(dual=False, max_iter=5000))\n",
    "modelo.fit(treino_x, treino_y)\n",
    "previsoes = modelo.predict(teste_x)\n",
    "acuracia = accuracy_score(teste_y, previsoes) * 100\n",
    "treino_y.value_counts()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparing to baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 66.53%\n",
      "The baseline accuracy using the type of school was 49.82%\n",
      "The random baseline accuracy was 50.16%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\" % acuracia)\n",
    "baseline_testing(teste_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Human Sciences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    346008\n",
       "0    345410\n",
       "Name: CH, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = brDataEnem['CH']\n",
    "\n",
    "#Defining the training set \n",
    "SEED = 30\n",
    "treino_x, teste_x, treino_y, teste_y = train_test_split(X, y, random_state = SEED, test_size = 0.25)\n",
    "\n",
    "#Traning\n",
    "modelo = make_pipeline(StandardScaler(),LinearSVC(dual=False, max_iter=5000))\n",
    "modelo.fit(treino_x, treino_y)\n",
    "previsoes = modelo.predict(teste_x)\n",
    "acuracia = accuracy_score(teste_y, previsoes) * 100\n",
    "treino_y.value_counts()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparing to baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 67.06%\n",
      "The baseline accuracy using the type of school was 49.83%\n",
      "The random baseline accuracy was 50.01%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\" % acuracia)\n",
    "baseline_testing(teste_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Writing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    350308\n",
       "0    341110\n",
       "Name: Re, dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = brDataEnem['Re']\n",
    "\n",
    "#Defining the training set \n",
    "SEED = 30\n",
    "treino_x, teste_x, treino_y, teste_y = train_test_split(X, y, random_state = SEED, test_size = 0.25)\n",
    "\n",
    "#Traning\n",
    "modelo = make_pipeline(StandardScaler(),LinearSVC(dual=False, max_iter=5000))\n",
    "modelo.fit(treino_x, treino_y)\n",
    "previsoes = modelo.predict(teste_x)\n",
    "acuracia = accuracy_score(teste_y, previsoes) * 100\n",
    "treino_y.value_counts()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Comparing to baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 63.84%\n",
      "The baseline accuracy using the type of school was 49.42%\n",
      "The random baseline accuracy was 50.66%\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: %.2f%%\" % acuracia)\n",
    "baseline_testing(teste_y)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Cópia de enemnotebook.ipynb",
   "provenance": [
    {
     "file_id": "1-968hmwfbH4B7qeB33JIGNmyzNSiFq4F",
     "timestamp": 1601917218967
    }
   ]
  },
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}