imbalance_ensemble.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_classification\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import balanced_accuracy_score, confusion_matrix\n",
    "from imblearn.ensemble import (BalancedBaggingClassifier, BalancedRandomForestClassifier,\n",
    "                               RUSBoostClassifier, EasyEnsembleClassifier)\n",
    "from imblearn.metrics import classification_report_imbalanced\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import BaggingClassifier, AdaBoostClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn import metrics, svm\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import itertools\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from mlxtend.plotting import plot_decision_regions\n",
    "from collections import Counter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados em cada uma das classes [(0, 63), (1, 262), (2, 4675)]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((5000, 6), (5000,))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X,y =make_classification(n_samples=5000, n_features=6, n_informative=2, n_classes=3, n_clusters_per_class=1,\n",
    "                        n_redundant=0,n_repeated=0,weights=[0.01,0.05,0.94],class_sep=0.8,random_state=0)\n",
    "print(\"Dados em cada uma das classes\",sorted(Counter(y).items()))\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# funcão usada para imprimir a confusion matrix\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "    else:\n",
    "        print('')\n",
    "\n",
    "\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modelo imbalanced"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Bagging Classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "O método bagging faz uso de vários estimadores aplicados a diferentes amostras selecionadas aleatóriamente, mas acaba priorizando a classe com maior número de dados. Veja abaixo que treinamos o nosso \"classificador\" e depois testamos o mesmo para saber se obteve um bom resultado. Nosso conjunto de teste tem o seguinte número de dados em cada um das classes: classe0: 21, classe1: 55 e classe2: 1174. Ao aplicar o \"classificador\" nesse conjunto de dados obtemos como resultado um score de 0.93. A confusion matrix mostra que erramos pouco, com destaque para a previsão do resultado da classe 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados para teste:  [(0, 21), (1, 55), (2, 1174)]\n",
      "Dados previstos:  [(0, 19), (1, 52), (2, 1179)]\n",
      "Score para BC usando Decision Tree: 0.9310385431339435\n"
     ]
    }
   ],
   "source": [
    "# Bagging Classifier\n",
    "x_train, x_test, y_train, y_test = train_test_split(X,y,random_state=0)\n",
    "print(\"Dados para teste: \",sorted(Counter(y_test).items()))\n",
    "bc=BaggingClassifier(base_estimator=DecisionTreeClassifier(),random_state=0)\n",
    "m1=bc.fit(x_train,y_train)\n",
    "y_pred1=m1.predict(x_test)\n",
    "print(\"Dados previstos: \",sorted(Counter(y_pred1).items()))\n",
    "print(\"Score para BC usando Decision Tree:\", balanced_accuracy_score(y_test,y_pred1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                   pre       rec       spe        f1       geo       iba       sup\n",
      "\n",
      "          0       1.00      0.90      1.00      0.95      0.95      0.90        21\n",
      "          1       0.94      0.89      1.00      0.92      0.94      0.88        55\n",
      "          2       0.99      1.00      0.89      1.00      0.94      0.90      1174\n",
      "\n",
      "avg / total       0.99      0.99      0.90      0.99      0.94      0.90      1250\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report_imbalanced(y_test,y_pred1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_tree = confusion_matrix(y_test, y_pred1)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm_tree, classes=[0,1,2],\n",
    "                      title='Usando Bagging Classifier sem correção')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Corrigindo o imbalanced"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora podemos tentar um outro classificador para fazer o rebalanceamento nas estimativas. Como primeiro exemplo, usamos o método \"BalancedBaggingClassifier\". Esse faz um \"resample\" de cada um dos subconjunto de dados da amostra para treinar cada um dos estimadores para depois combinar os resultados. Nesse caso, veja a diferença de classificação da classe0 e classe1 em relação ao primeiro método."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Balanced Bagging Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados para teste:  [(0, 21), (1, 55), (2, 1174)]\n",
      "Dados rebalanceados:  [(0, 61), (1, 168), (2, 1021)]\n",
      "Score para BBC usando Decision Tree: 0.898866739922958\n"
     ]
    }
   ],
   "source": [
    "bbc=BalancedBaggingClassifier(base_estimator=DecisionTreeClassifier(),sampling_strategy='auto',\n",
    "                             replacement=False,random_state=0)\n",
    "m2=bbc.fit(x_train,y_train)\n",
    "y_pred2=m2.predict(x_test)\n",
    "print(\"Dados para teste: \",sorted(Counter(y_test).items()))\n",
    "print(\"Dados rebalanceados: \",sorted(Counter(y_pred2).items()))\n",
    "print(\"Score para BBC usando Decision Tree:\", balanced_accuracy_score(y_test,y_pred2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                   pre       rec       spe        f1       geo       iba       sup\n",
      "\n",
      "          0       0.31      0.90      0.97      0.46      0.93      0.87        21\n",
      "          1       0.30      0.93      0.90      0.46      0.91      0.84        55\n",
      "          2       0.99      0.86      0.92      0.92      0.89      0.79      1174\n",
      "\n",
      "avg / total       0.95      0.87      0.92      0.90      0.89      0.80      1250\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report_imbalanced(y_test,y_pred2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_tree2 = confusion_matrix(y_test, y_pred2)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm_tree2, classes=[0,1,2],\n",
    "                      title='Usando BalancedBaggingClassifier')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Forest of Randomized Trees\n",
    "No método ensemble usado a seguir cada uma das árvores da floresta tem uma amostra que é balanceada via bootstrap."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados para teste:  [(0, 21), (1, 55), (2, 1174)]\n",
      "Dados rebalanceados:  [(0, 38), (1, 163), (2, 1049)]\n",
      "Score para BC usando Decision Tree: 0.8162289726173882\n"
     ]
    }
   ],
   "source": [
    "brf=BalancedRandomForestClassifier(n_estimators=100,random_state=0,sampling_strategy='auto', \n",
    "                                   replacement=False, n_jobs=1,warm_start=True)\n",
    "m3=brf.fit(x_train,y_train)\n",
    "y_pred3=m3.predict(x_test)\n",
    "print(\"Dados para teste: \",sorted(Counter(y_test).items()))\n",
    "print(\"Dados rebalanceados: \",sorted(Counter(y_pred3).items()))\n",
    "print(\"Score para BC usando Decision Tree:\", balanced_accuracy_score(y_test,y_pred3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_tree3 = confusion_matrix(y_test, y_pred3)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm_tree3, classes=[0,1,2],\n",
    "                      title='Usando BalancedRandomForestClassifier')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. RUSBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados para teste:  [(0, 21), (1, 55), (2, 1174)]\n",
      "Dados rebalanceados:  [(0, 18), (1, 203), (2, 1029)]\n",
      "Score para BC usando Decision Tree: 0.7133247785717973\n"
     ]
    }
   ],
   "source": [
    "rus=RUSBoostClassifier(base_estimator=DecisionTreeClassifier(), n_estimators=100, learning_rate=1.0, \n",
    "                       algorithm='SAMME.R', sampling_strategy='auto', replacement=False, random_state=0)\n",
    "m4=rus.fit(x_train,y_train)\n",
    "y_pred4=m4.predict(x_test)\n",
    "print(\"Dados para teste: \",sorted(Counter(y_test).items()))\n",
    "print(\"Dados rebalanceados: \",sorted(Counter(y_pred4).items()))\n",
    "print(\"Score para BC usando Decision Tree:\", balanced_accuracy_score(y_test,y_pred4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_tree4 = confusion_matrix(y_test, y_pred4)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm_tree4, classes=[0,1,2],\n",
    "                      title='Usando RUSBoostClassifier')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. EasyEnsembleClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dados para teste:  [(0, 21), (1, 55), (2, 1174)]\n",
      "Dados rebalanceados:  [(0, 391), (1, 83), (2, 776)]\n",
      "Score para BC usando Decision Tree: 0.7937921438773228\n"
     ]
    }
   ],
   "source": [
    "eec=EasyEnsembleClassifier(n_estimators=100, base_estimator=AdaBoostClassifier(), warm_start=True, \n",
    "                           sampling_strategy='auto', replacement=False, n_jobs=1, random_state=0, verbose=0)\n",
    "m5=eec.fit(x_train,y_train)\n",
    "y_pred5=m5.predict(x_test)\n",
    "print(\"Dados para teste: \",sorted(Counter(y_test).items()))\n",
    "print(\"Dados rebalanceados: \",sorted(Counter(y_pred5).items()))\n",
    "print(\"Score para BC usando Decision Tree:\", balanced_accuracy_score(y_test,y_pred5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm_tree5 = confusion_matrix(y_test, y_pred5)\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cm_tree5, classes=[0,1,2],\n",
    "                      title='Usando EasyEnsembleClassifier')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Uma boa discussão sobre o tema pode ser vista em:\n",
    "https://www.analyticsvidhya.com/blog/2018/06/comprehensive-guide-for-ensemble-models/ "
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}