RegLogistica_Hands_On.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Vinicius F. Caridá</b>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1 style=\"text-align:center\"> Machine Learning Hands-On </h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3 class=\"title\"> Links úteis: </h3>\n",
    "<br>\n",
    "<ol>\n",
    "  <li class=\"item\"><a href = \"http://scikit-learn.org/stable/\"> Sklearn</a>: principal biblioteca de aprendizado de máquina para python.</li>\n",
    "  <li class=\"item\"><a href = \"http://pandas.pydata.org/pandas-docs/stable/\"> Pandas</a>: (quase) tudo o que você precisa para trabalhar rapidamente com tabelas</li>\n",
    "  <li class=\"item\"><a href = \"https://docs.scipy.org/doc/numpy/reference/index.html\"> Numpy</a>: funções matemáticas estão aqui</li>\n",
    "  <li class=\"item\"><a href = \"https://matplotlib.org/contents.html\"> Matplotlib</a>: te ajuda a plotar!</li>\n",
    "  <li class=\"item\"><a href = \"https://seaborn.pydata.org/api.html\"> Seaborn</a>: Deixa a matplotlib bonita (contém plots mais bem estruturados)</li>\n",
    "</ol>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Trabalharemos com as seguintes bases de dados:</b>\n",
    "- iris.csv\n",
    "- lenses_final.csv\n",
    "- car.data\n",
    "- horse-colic.data\n",
    "- PETR4.SA.csv\n",
    "- Boston.csv\n",
    "- artificial_data\n",
    "\n",
    "<b>Testem a leitura de um dos arquivos deste curso</b>\n",
    "- Acessem a documentação da função <i>read_csv</i> do <i>pandas</i>: https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_csv.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_style('whitegrid', {\"axes.grid\" : False})\n",
    "sns.set_context('notebook')\n",
    "np.random.seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\vinicius\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:4: FutureWarning: read_table is deprecated, use read_csv instead, passing sep='\\t'.\n",
      "  after removing the cwd from sys.path.\n"
     ]
    }
   ],
   "source": [
    "iris            = pd.read_csv(\"./data/iris.csv\")\n",
    "lenses          = pd.read_csv(\"./data/lenses_final.csv\")\n",
    "car             = pd.read_csv(\"./data/cars.csv\")\n",
    "horsecolic      = pd.read_table(\"./data/horse-colic.data\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2 style=\"text-align:center; margin:40px\"> Regressão Logística </h2>\n",
    "\n",
    "![](img/logreg.png)\n",
    "\n",
    "A <b>regressão logística</b> é um tipo de regressão utilizada no caso de resposta binária. O valor retornado pela regressão logística representa a probabilidade de ocorrer o evento de interesse.\n",
    "\n",
    "Assim como na regressão linear, queremos encontrar os parâmetros $\\beta$ que fazem o modelo melhor se ajustar aos dados. Partimos de uma função linear:\n",
    "\n",
    "$$ f_{\\beta}(x) = \\sum_{i = 0}^{N} \\beta_ix_i$$\n",
    "\n",
    "E \"encapsulamos\" a função linear com uma função (chamada de <b>logística</b> ou <b>sigmoide</b>) que mapeia o espaço dos números reais em um espaço reduzido entre 0 e 1:\n",
    "\n",
    "$$ \\sigma(z) = \\frac{1}{1 + e^{-z}}$$\n",
    "\n",
    "<b>O novo modelo é dado por:</b>\n",
    "\n",
    "$$ f_{\\beta}(x) = \\sigma(\\sum_{i = 0}^{N} \\beta_ix_i) = \\frac{1}{1 + e^{-\\sum_{i = 0}^{N} \\beta_ix_i}}$$\n",
    "\n",
    "Também substituímos a <b>função de erro</b> por uma função mais adequada para a variável resposta binária:\n",
    "\n",
    "$$ L(x_i, y_i) = \\color{red}{-y_i\\log(f_{\\beta}(x_i))}\\color{blue}{-(1-y_i)\\log(1-f_{\\beta}(x_i))}$$\n",
    "\n",
    "E mais uma vez aplicamos o método de gradiente descendente para encontrar o valor ótimo de $\\beta$."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3> Scikit-Learn </h3>\n",
    "Agora, vamos rodar os códigos abaixo que usam uma implementação do algoritmo disponível no pacote <i>sklearn</i> (Scikit-Learn). Mas antes vamos dar uma olhada na base que usaremos com os próximos algoritmos de classificação"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x20d3fb84320>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "baseDados = pd.read_csv('data/base_regressao_logistica.csv')\n",
    "plt.scatter(baseDados.X1[baseDados.Y == 0], baseDados.X2[baseDados.Y == 0], c = 'darkgreen', marker = '^')\n",
    "plt.scatter(baseDados.X1[baseDados.Y == 1], baseDados.X2[baseDados.Y == 1], c = 'black', marker = ',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.preprocessing import PolynomialFeatures"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No regressor logístico do _sklearn_ utilizamos um parâmetro de regularização $ C $.\n",
    "\n",
    "A fórmula de custo $J(f_{\\beta}, y)$ recebe um termo adicional $\\frac{1}{C} \\sum_{i=1}^{n} \\beta_i^2$ (regularização $l_2$).\n",
    "\n",
    "O parâmetro $ C $ controla quão grandes os coeficientes $ \\beta_i $ podem se tornar. Quanto menor o valor de $ C $, mais penalizado o modelo será por utilizar os coeficientes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\vinicius\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LR = LogisticRegression(C = 1)\n",
    "X = baseDados.loc[:, baseDados.columns != 'Y']\n",
    "Y = baseDados.Y\n",
    "LR.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x20d40709198>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_min, x_max = baseDados.X1.min() - .1, baseDados.X1.max() + .1\n",
    "y_min, y_max = baseDados.X2.min() - .1, baseDados.X2.max() + .1\n",
    "h = .005\n",
    "xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
    "Z = LR.predict(np.c_[xx.ravel(), yy.ravel()])\n",
    "\n",
    "Z = Z.reshape(xx.shape)\n",
    "plt.figure()\n",
    "plt.pcolormesh(xx, yy, Z, cmap = plt.cm.Accent)\n",
    "\n",
    "pred = LR.predict(X)\n",
    "plt.scatter(baseDados.X1[Y == 0], baseDados.X2[Y == 0], c = 'darkgreen', marker = '^')\n",
    "plt.scatter(baseDados.X1[Y == 1], baseDados.X2[Y == 1], c = 'black', marker = ',')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora iremos treinar um modelo mais complexo. Agora teremos atributos polinomiais, e então, além de ter coeficientes para as variáveis preditivas, também teremos para suas combinações polinomiais, como por exemplo $ x_1^2, x_2^2, x_1x_2, x_1^2x_2, x_1x_2^2, x_1^3, x_2^3 $. Por exemplo, para atributos quadráticos temos $ f_{\\beta}(x) = {\\beta_0} + {\\beta_1}x_1 + {\\beta_2}x_1^2 + {\\beta_3}x_2 + {\\beta_4}x_2^2 + {\\beta_5}x_1x_2 $"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\vinicius\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=10000000000.0, class_weight=None, dual=False,\n",
       "          fit_intercept=True, intercept_scaling=1, max_iter=100,\n",
       "          multi_class='warn', n_jobs=None, penalty='l2', random_state=None,\n",
       "          solver='warn', tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "LR = LogisticRegression(C = 1e10)\n",
    "grau = 10\n",
    "X = PolynomialFeatures(degree = grau).fit_transform(baseDados[baseDados.columns[:-1]])\n",
    "Y = baseDados['Y']\n",
    "LR.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x20d4090b550>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x_min, x_max = baseDados.X1.min() - .1, baseDados.X1.max() + .1\n",
    "y_min, y_max = baseDados.X2.min() - .1, baseDados.X2.max() + .1\n",
    "h = .005\n",
    "xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
    "Z = LR.predict(PolynomialFeatures(degree = grau).fit_transform(np.c_[xx.ravel(),yy.ravel()]))\n",
    "\n",
    "Z = Z.reshape(xx.shape)\n",
    "plt.figure()\n",
    "plt.pcolormesh(xx, yy, Z, cmap = plt.cm.Accent)\n",
    "\n",
    "pred = LR.predict(X)\n",
    "plt.scatter(baseDados.X1[Y == 0], baseDados.X2[Y == 0], c = 'darkgreen', marker = '^')\n",
    "plt.scatter(baseDados.X1[Y == 1], baseDados.X2[Y == 1], c = 'black', marker = ',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos utilizar os mesmos dados empregados no modelo de regressão linear para uma outra tarefa. Classificar o indivíduo nos sexos masculino e feminino dado o seu peso e altura. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import mean_squared_error\n",
    "%matplotlib inline\n",
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "pesos = pd.read_csv('weights.csv', sep=';')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Como já fizemos uma análise exploratória de dados no exercício de regressão linear, vamos pular essa parte e seguirmos direto para a remoção dos outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x20d409226a0>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pesos = pesos[pesos.weight <= 100].reset_index(drop=True)\n",
    "plt.scatter(x=pesos.loc[pesos.sex=='M', 'height'], y=pesos.loc[pesos.sex=='M', 'weight'], c='blue')\n",
    "plt.scatter(x=pesos.loc[pesos.sex=='F', 'height'], y=pesos.loc[pesos.sex=='F', 'weight'], c='red')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Observe que, apesar de uma leve sobreposição, conseguimos separar os indivídos do sexo masculino com uma reta. O que vamos encontrar com nossa Regressão Logística é a reta que melhor separa esses dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>sex</th>\n",
       "      <th>weight</th>\n",
       "      <th>height</th>\n",
       "      <th>M</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>M</td>\n",
       "      <td>77</td>\n",
       "      <td>182</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>F</td>\n",
       "      <td>58</td>\n",
       "      <td>161</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>F</td>\n",
       "      <td>53</td>\n",
       "      <td>161</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>M</td>\n",
       "      <td>68</td>\n",
       "      <td>177</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>F</td>\n",
       "      <td>59</td>\n",
       "      <td>157</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  sex  weight  height  M\n",
       "0   M      77     182  1\n",
       "1   F      58     161  0\n",
       "2   F      53     161  0\n",
       "3   M      68     177  1\n",
       "4   F      59     157  0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummies_sex = pd.get_dummies(pesos.sex, drop_first=True)\n",
    "pesos = pd.concat([pesos, dummies_sex], axis=1)\n",
    "pesos.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pesos[['weight', 'height']]\n",
    "y = pesos.M"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\vinicius\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:433: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='warn',\n",
       "          n_jobs=None, penalty='l2', random_state=None, solver='warn',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg_log = LogisticRegression()\n",
    "reg_log.fit(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Assim como na regressão linear, podemos observar os coeficientes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.21993397, -0.07531506]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg_log.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-1.55761751])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg_log.intercept_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Isso quer dizer que a equação da nossa função logística é: <br>\n",
    "    $prob(homem) = \\frac{1}{1 + e^{-(peso\\times 0.2199 - altura\\times 0.0753 - 1.5576)}}$ "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Com esses coeficientes identificamos que a reta que separa indivíduos masculinos de femininos é: <br>\n",
    "$peso\\times 0.2199 - altura\\times 0.0753 - 1.5576 = 0$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Podemos visualizar essa reta, para isso devemos escrevê-la em função de $peso$ ou $altura$, já que meu $y$ é fixo em $0$. <br>\n",
    "Com algumas manipulações algébricas temos a reta em função da $altura$ dada por: <br><br>\n",
    "    $peso = \\frac{-altura\\times 0.0753 + 1.5576}{0.2199}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "heights = np.linspace(140, 210, 70)\n",
    "coef_weight = float(reg_log.coef_[0][0])\n",
    "coef_height = float(reg_log.coef_[0][1])\n",
    "intercept = float(reg_log.intercept_)\n",
    "\n",
    "reta_decisao = (-heights*coef_height - intercept) / coef_weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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": [
    "plt.scatter(x=pesos.loc[pesos.sex=='M', 'height'], y=pesos.loc[pesos.sex=='M', 'weight'], c='blue')\n",
    "plt.scatter(x=pesos.loc[pesos.sex=='F', 'height'], y=pesos.loc[pesos.sex=='F', 'weight'], c='red')\n",
    "plt.plot(heights, reta_decisao, '-', c='black')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Vamos visualizar os dados classificados conferirmos se foram classificados de acordo com nossa superfície de decisão. Para isso classificamos esses mesmos dados, assim obtemos os rótulos dados durante o treinamento."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1,\n",
       "       0, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1,\n",
       "       0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "       1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1,\n",
       "       0, 1, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0,\n",
       "       0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1,\n",
       "       0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1],\n",
       "      dtype=uint8)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classified_reg = reg_log.predict(X)\n",
    "classified_reg"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Concatenamos esses dados (na forma de um dataframe), aos dados originais para conseguirmos plotá-los."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "pesos_classified_reg = pd.concat([pesos, pd.DataFrame(classified_reg, columns=['classificado_reg'])], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pesos_classified_reg.head(100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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": [
    "plt.scatter(x=pesos_classified_reg.loc[pesos_classified_reg.classificado_reg==1.0, 'height'], y=pesos_classified_reg.loc[pesos_classified_reg.classificado_reg==1.0, 'weight'], c='blue')\n",
    "plt.scatter(x=pesos_classified_reg.loc[pesos_classified_reg.classificado_reg==0.0, 'height'], y=pesos_classified_reg.loc[pesos_classified_reg.classificado_reg==0.0, 'weight'], c='red')\n",
    "plt.plot(heights, reta_decisao, '-', c='black')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Como esperado, os indivíduos \"acima\" da reta tem valor positivo em relação a ela, logo assumem a classe 1 (masculino). Já os \"abaixo\", possuem valor negativo e assumem a classe 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Como o modelo de Regressão Logística é probabilístico, muitas vezes é interessante obtermos a probabilidade de um elemento pertencer à classe determinada. Neste caso usamos o método predict_proba()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "novo_exemplo = [[75.00, 173]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.12921413, 0.87078587]])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reg_log.predict_proba(novo_exemplo)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.86      0.83       111\n",
      "           1       0.80      0.71      0.75        84\n",
      "\n",
      "   micro avg       0.80      0.80      0.80       195\n",
      "   macro avg       0.80      0.79      0.79       195\n",
      "weighted avg       0.80      0.80      0.80       195\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "confusion_matrix = confusion_matrix(y, classified_reg)\n",
    "confusion_matrix\n",
    "print(classification_report(y, classified_reg))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Portanto, para nosso modelo, uma pessoa de 1,73m e 75Kg, a probabilidade de ser do sexo masculino é 87% e do feminino 13%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>weight</th>\n",
       "      <th>height</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>77</td>\n",
       "      <td>182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>58</td>\n",
       "      <td>161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>53</td>\n",
       "      <td>161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>68</td>\n",
       "      <td>177</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>59</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   weight  height\n",
       "0      77     182\n",
       "1      58     161\n",
       "2      53     161\n",
       "3      68     177\n",
       "4      59     157"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Em um plano 3D conseguimer ver a sigmoid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from mpl_toolkits.mplot3d import Axes3D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "weights = X['weight']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "heights = X['height']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "probs = pd.DataFrame(reg_log.predict_proba(X)[:,1], columns=['probs'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "pesos_classificados = pd.concat([X, probs], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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>weight</th>\n",
       "      <th>height</th>\n",
       "      <th>probs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>77</td>\n",
       "      <td>182</td>\n",
       "      <td>0.841570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>58</td>\n",
       "      <td>161</td>\n",
       "      <td>0.283499</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>53</td>\n",
       "      <td>161</td>\n",
       "      <td>0.116413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>68</td>\n",
       "      <td>177</td>\n",
       "      <td>0.516775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>59</td>\n",
       "      <td>157</td>\n",
       "      <td>0.399877</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   weight  height     probs\n",
       "0      77     182  0.841570\n",
       "1      58     161  0.283499\n",
       "2      53     161  0.116413\n",
       "3      68     177  0.516775\n",
       "4      59     157  0.399877"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pesos_classificados.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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": [
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111, projection='3d')\n",
    "ax.scatter(xs=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'height'],\n",
    "           ys=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'weight'], \n",
    "           zs=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'probs'],\n",
    "           c='blue', marker='o')\n",
    "ax.scatter(xs=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'height'],\n",
    "           ys=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'weight'], \n",
    "           zs=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'probs'],\n",
    "           c='red', marker='o')\n",
    "ax.azim = 350 #alterar o valore aqui para rotacionar o grafico\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Podemos ver a superfície de decisão"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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": [
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111, projection='3d')\n",
    "ax.scatter(xs=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'height'],\n",
    "           ys=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'weight'], \n",
    "           zs=pesos_classificados.loc[pesos_classificados.probs >= 0.5, 'probs'],\n",
    "           c='blue', marker='o')\n",
    "ax.scatter(xs=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'height'],\n",
    "           ys=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'weight'], \n",
    "           zs=pesos_classificados.loc[pesos_classificados.probs < 0.5, 'probs'],\n",
    "           c='red', marker='o')\n",
    "ax.scatter(xs=np.linspace(140, 210, 70),\n",
    "           ys=np.linspace(40, 90, 70), \n",
    "           zs=np.ones(70)*0.5, marker='_', c='black')\n",
    "ax.azim = 350"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}