Predictive Statistics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning Project - Predictive Statistics - 50pts\n",
    "\n",
    "Kathia Zamora Arias <kattia.zamora@prodigious.com><br />\n",
    "Alejandro Madrigal Leival <me@alemadlei.tech>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn import decomposition\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report\n",
    "from scipy.spatial.distance import cdist\n",
    "from sklearn.model_selection import cross_val_score, cross_val_predict, GridSearchCV\n",
    "from sklearn import metrics\n",
    "from sklearn.preprocessing import Normalizer, RobustScaler, MinMaxScaler, StandardScaler\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "import numpy as np\n",
    "from time import time\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>ID</th>\n",
       "      <th>LIMIT_BAL</th>\n",
       "      <th>SEX</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>MARRIAGE</th>\n",
       "      <th>AGE</th>\n",
       "      <th>PAY_0</th>\n",
       "      <th>PAY_2</th>\n",
       "      <th>PAY_3</th>\n",
       "      <th>PAY_4</th>\n",
       "      <th>...</th>\n",
       "      <th>BILL_AMT4</th>\n",
       "      <th>BILL_AMT5</th>\n",
       "      <th>BILL_AMT6</th>\n",
       "      <th>PAY_AMT1</th>\n",
       "      <th>PAY_AMT2</th>\n",
       "      <th>PAY_AMT3</th>\n",
       "      <th>PAY_AMT4</th>\n",
       "      <th>PAY_AMT5</th>\n",
       "      <th>PAY_AMT6</th>\n",
       "      <th>default payment next month</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>20000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>689</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>120000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>26</td>\n",
       "      <td>-1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>3272</td>\n",
       "      <td>3455</td>\n",
       "      <td>3261</td>\n",
       "      <td>0</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>0</td>\n",
       "      <td>2000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>90000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>34</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>14331</td>\n",
       "      <td>14948</td>\n",
       "      <td>15549</td>\n",
       "      <td>1518</td>\n",
       "      <td>1500</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>5000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>50000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>37</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>28314</td>\n",
       "      <td>28959</td>\n",
       "      <td>29547</td>\n",
       "      <td>2000</td>\n",
       "      <td>2019</td>\n",
       "      <td>1200</td>\n",
       "      <td>1100</td>\n",
       "      <td>1069</td>\n",
       "      <td>1000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>50000</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>57</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>20940</td>\n",
       "      <td>19146</td>\n",
       "      <td>19131</td>\n",
       "      <td>2000</td>\n",
       "      <td>36681</td>\n",
       "      <td>10000</td>\n",
       "      <td>9000</td>\n",
       "      <td>689</td>\n",
       "      <td>679</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID  LIMIT_BAL  SEX  EDUCATION  MARRIAGE  AGE  PAY_0  PAY_2  PAY_3  PAY_4  \\\n",
       "0   1      20000    2          2         1   24      2      2     -1     -1   \n",
       "1   2     120000    2          2         2   26     -1      2      0      0   \n",
       "2   3      90000    2          2         2   34      0      0      0      0   \n",
       "3   4      50000    2          2         1   37      0      0      0      0   \n",
       "4   5      50000    1          2         1   57     -1      0     -1      0   \n",
       "\n",
       "   ...  BILL_AMT4  BILL_AMT5  BILL_AMT6  PAY_AMT1  PAY_AMT2  PAY_AMT3  \\\n",
       "0  ...          0          0          0         0       689         0   \n",
       "1  ...       3272       3455       3261         0      1000      1000   \n",
       "2  ...      14331      14948      15549      1518      1500      1000   \n",
       "3  ...      28314      28959      29547      2000      2019      1200   \n",
       "4  ...      20940      19146      19131      2000     36681     10000   \n",
       "\n",
       "   PAY_AMT4  PAY_AMT5  PAY_AMT6  default payment next month  \n",
       "0         0         0         0                           1  \n",
       "1      1000         0      2000                           1  \n",
       "2      1000      1000      5000                           0  \n",
       "3      1100      1069      1000                           0  \n",
       "4      9000       689       679                           0  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Description\n",
    "# LIMIT_BAL: Credit in Taiwan dollars.\n",
    "# SEX: \n",
    "#   1 => Male\n",
    "#   2 => Female\n",
    "# EDUCATION: \n",
    "#   1 => Graduate School\n",
    "#   2 => University\n",
    "#   3 => High School\n",
    "#   4 => Others\n",
    "# MARRIAGE:\n",
    "#   1 => Married\n",
    "#   2 => Single\n",
    "#   3 => Other\n",
    "# AGE: Age\n",
    "# PAY_0: Payment status for September 2005\n",
    "# PAY_2: Payment status for August 2005\n",
    "# PAY_3: Payment status for July 2005\n",
    "# PAY_4: Payment status for June 2005\n",
    "# PAY_5: Payment status for May 2005\n",
    "# PAY_6: Payment status for April 2005\n",
    "# - All of these have the following possible values\n",
    "#   -1 => Pay duly\n",
    "#    1 => One month payment delay\n",
    "#    2 => Two months payment delay\n",
    "#    . => . months payment delay\n",
    "#    9 => Nine months or month payment delay.\n",
    "# BILL_AMT1: Bill statement amount for September 2005.\n",
    "# BILL_AMT2: Bill statement amount for August 2005.\n",
    "# BILL_AMT3: Bill statement amount for July 2005.\n",
    "# BILL_AMT4: Bill statement amount for June 2005.\n",
    "# BILL_AMT5: Bill statement amount for May 2005.\n",
    "# BILL_AMT6: Bill statement amount for April 2005.\n",
    "# - All of these values are in Taiwan dollars.\n",
    "# PAY_AMT1: Amount of previous payment for September 2005.\n",
    "# PAY_AMT2: Amount of previous payment for August 2005.\n",
    "# PAY_AMT3: Amount of previous payment for July 2005.\n",
    "# PAY_AMT4: Amount of previous payment for June 2005.\n",
    "# PAY_AMT5: Amount of previous payment for May 2005.\n",
    "# PAY_AMT6: Amount of previous payment for April 2005.\n",
    "# default payment next month: Indicates if the next month (October 2005) will be a default payment month.\n",
    "\n",
    "# Loads and cleans up our dataset (first row in this case is the second one).\n",
    "data = pd.read_csv('default_credit.csv', header=1);\n",
    "data = data.astype('int32')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1 - Can we cluster our customers into categories? (unlabeled clustering) - (10 pts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# As our data is not labeled in any way that would allows us to classify the people, we will attempt to\n",
    "# figure this out by using KMeans.\n",
    "\n",
    "# We get the data use to infere the clusters.\n",
    "y = data['default payment next month'];\n",
    "X = data.drop(['default payment next month', 'ID'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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": [
    "# Uses PCA to transform data.\n",
    "pca = decomposition.PCA(n_components = 2)\n",
    "pca.fit(X)\n",
    "X = pca.transform(X)\n",
    "\n",
    "# Gets the clusters groups to plot and figure out the best number of clusters to use.\n",
    "distortions = []\n",
    "K = range(1,10)\n",
    "for k in K:\n",
    "    kmeanModel = KMeans(n_clusters=k).fit(X)\n",
    "    distortions.append(sum(np.min(cdist(X, kmeanModel.cluster_centers_, 'euclidean'), axis=1)) / X.shape[0])\n",
    "\n",
    "# Plot the elbow\n",
    "plt.plot(K, distortions, 'bx-')\n",
    "plt.xlabel('k')\n",
    "plt.ylabel('Distortion')\n",
    "plt.title('The Elbow Method showing the optimal k')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "# Lets compare the groupings that could be used as elbows.\n",
    "for i in range(3, 6):\n",
    "    kmeans = KMeans(n_clusters=i)\n",
    "    kmeans.fit(X)\n",
    "    predict = kmeans.predict(X)\n",
    "    centroids = kmeans.cluster_centers_\n",
    "    \n",
    "    plt.title('With ' + str(i) + ' clusters')\n",
    "    plt.scatter(X[ : , 0], X[ : , 1], s=50, c=predict)\n",
    "    plt.scatter(centroids[ : , 0], centroids[ : , 1], s = 200, marker='x', c = 'm')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Based on the data, we feel that 5 clusters is something that resembles the data we have with a good accuracy."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2 - Can cluster categories serve to predict payment defaults? – (10 pts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluation scores for 3 clusters.\n",
      "Cross-validated scores: [1.17333333 1.44966667 1.44566667 0.75533333 2.08766667 1.25766667\n",
      " 2.07633333 2.01333333 1.382      2.02633333]\n",
      "Cross-Predicted Accuracy: 6.678232039956992\n",
      "================================\n",
      "Evaluation scores for 4 clusters.\n",
      "Cross-validated scores: [1.65066667 2.554      1.11666667 1.082      1.76833333 5.20933333\n",
      " 3.07833333 2.41966667 4.215      1.55566667]\n",
      "Cross-Predicted Accuracy: 14.429991829906012\n",
      "================================\n",
      "Evaluation scores for 5 clusters.\n",
      "Cross-validated scores: [3.32       5.28633333 5.89333333 5.033      5.15766667 4.033\n",
      " 4.9        5.92633333 4.84333333 7.687     ]\n",
      "Cross-Predicted Accuracy: 29.129543511864902\n",
      "================================\n"
     ]
    }
   ],
   "source": [
    "# Will iterate over cluster variations...\n",
    "for i in range(3, 6):\n",
    "\n",
    "    print ('Evaluation scores for', str(i), 'clusters.')\n",
    "    kmeans_model_cross_val = KMeans(n_clusters=i)\n",
    "\n",
    "    # Cross validates data \n",
    "    scores = cross_val_score(kmeans_model_cross_val, X, y, scoring='neg_mean_squared_error', cv=10)\n",
    "    print('Cross-validated scores:', -scores)\n",
    "\n",
    "    predictions = cross_val_predict(kmeans_model_cross_val, X, y, cv=10)\n",
    "    accuracy = metrics.r2_score(y, predictions)\n",
    "    print('Cross-Predicted Accuracy:', -accuracy)\n",
    "    print('================================')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Based on the data, our initial assumption that 5 clusters better represents the data seems to be valid, as it is the one with the highest score. However, because it is so low, we can confirm that using clusters to attempt to predict payment default is not going to provide any help at all."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3 - Can a classification algorithm be used to predict payment defaults based on the default data? (try as many ML algorithms as needed used in the class) - (25 pts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Before we do anything, we will clean up the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Clear all attributes that are not helpful to identify default payments.\n",
    "columns_to_remove = ['ID', 'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT3', 'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6']\n",
    "evaluation_X = data.drop(columns_to_remove, axis=1)\n",
    "\n",
    "# Result.\n",
    "evaluation_y = data['default payment next month'];\n",
    "\n",
    "# Change invalid marriage values to \"Other\"\n",
    "evaluation_X.loc[evaluation_X.MARRIAGE == 0, 'MARRIAGE'] = 3\n",
    "\n",
    "# Change invalid education values to \"Other\"\n",
    "evaluation_X.loc[evaluation_X.EDUCATION == 0, 'EDUCATION'] = 4\n",
    "evaluation_X.loc[evaluation_X.EDUCATION == 5, 'EDUCATION'] = 4\n",
    "evaluation_X.loc[evaluation_X.EDUCATION == 6, 'EDUCATION'] = 4\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Gets train and test data.\n",
    "X_train, X_test, y_train, y_test = train_test_split(evaluation_X, y, train_size=0.65, random_state=10)\n",
    "\n",
    "# Transforms the data with a normalizer.\n",
    "normd = Normalizer().fit(X_train)\n",
    "X_norm_train = normd.transform(X_train)\n",
    "X_norm_test = normd.transform(X_test)\n",
    "X_norm = normd.transform(evaluation_X)\n",
    "X_norm_train = pd.DataFrame(X_norm_train, columns=list(evaluation_X))\n",
    "X_norm_test = pd.DataFrame(X_norm_test, columns=list(evaluation_X))\n",
    "X_norm = pd.DataFrame(X_norm, columns=list(evaluation_X))\n",
    "\n",
    "# Transforms the data with a scaler.\n",
    "scaled = MinMaxScaler().fit(X_train)\n",
    "X_scaled_train = scaled.transform(X_train)\n",
    "X_scaled_test = scaled.transform(X_test)\n",
    "X_scaled = scaled.transform(evaluation_X)\n",
    "X_scaled_train = pd.DataFrame(X_scaled_train, columns=list(evaluation_X))\n",
    "X_scaled_test = pd.DataFrame(X_scaled_test, columns=list(evaluation_X))\n",
    "X_scaled = pd.DataFrame(X_scaled, columns=list(evaluation_X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Uses PCA to transform data. We use this because there are no category labels available.\n",
    "pca = decomposition.PCA(n_components = 2)\n",
    "pca.fit(X_train)\n",
    "X_train = pca.transform(X_train)\n",
    "X_test = pca.transform(X_test)\n",
    "\n",
    "pca_norm = decomposition.PCA(n_components = 2)\n",
    "pca_norm.fit(X_norm_train)\n",
    "X_norm_train = pca_norm.transform(X_norm_train)\n",
    "X_norm_test = pca_norm.transform(X_norm_test)\n",
    "\n",
    "pca_scaled = decomposition.PCA(n_components = 2)\n",
    "pca_scaled.fit(X_scaled_train)\n",
    "X_scaled_train = pca_scaled.transform(X_scaled_train)\n",
    "X_scaled_test = pca_scaled.transform(X_scaled_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Logistic regression evaluation\n",
    "## ==================================================================="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "param_grid = {\n",
    "  'random_state': [None, 2, 10],\n",
    "  'max_iter': [150], \n",
    "  'penalty': ['l2'],\n",
    "  'C': [1e3, 1e4, 5e4, 1e5],\n",
    "  'solver': ['lbfgs', 'sag']\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 12.527s:\n",
      "LogisticRegression(C=10000.0, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=150,\n",
      "                   multi_class='warn', n_jobs=-1, penalty='l2',\n",
      "                   random_state=None, solver='sag', tol=0.0001, verbose=0,\n",
      "                   warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.44      0.58      8134\n",
      "           1       0.28      0.74      0.40      2366\n",
      "\n",
      "    accuracy                           0.51     10500\n",
      "   macro avg       0.57      0.59      0.49     10500\n",
      "weighted avg       0.72      0.51      0.54     10500\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with logistic regression to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(LogisticRegression(n_jobs=-1), param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 10.519s:\n",
      "LogisticRegression(C=1000.0, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=150,\n",
      "                   multi_class='warn', n_jobs=-1, penalty='l2',\n",
      "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
      "                   warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      1.00      0.87      8134\n",
      "           1       0.00      0.00      0.00      2366\n",
      "\n",
      "    accuracy                           0.77     10500\n",
      "   macro avg       0.39      0.50      0.44     10500\n",
      "weighted avg       0.60      0.77      0.68     10500\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/alemadlei/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with logistic regression to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(LogisticRegression(n_jobs=-1), param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_norm_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_norm_pred = gs_model.predict(X_norm_test)\n",
    "print(classification_report(y_test, y_norm_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 26.602s:\n",
      "LogisticRegression(C=1000.0, class_weight=None, dual=False, fit_intercept=True,\n",
      "                   intercept_scaling=1, l1_ratio=None, max_iter=150,\n",
      "                   multi_class='warn', n_jobs=-1, penalty='l2',\n",
      "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
      "                   warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00      8134\n",
      "           1       1.00      1.00      1.00      2366\n",
      "\n",
      "    accuracy                           1.00     10500\n",
      "   macro avg       1.00      1.00      1.00     10500\n",
      "weighted avg       1.00      1.00      1.00     10500\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with logistic regression to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(LogisticRegression(n_jobs=-1), param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_scaled_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_scaled_pred = gs_model.predict(X_scaled_test)\n",
    "print(classification_report(y_test, y_scaled_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Though this results are very positive, it could be that the model is \"too good to be true\", and some overfitting might be happening. We test with the specific instance, and by using all data.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.82      0.97      0.89     23364\n",
      "           1       0.72      0.26      0.38      6636\n",
      "\n",
      "    accuracy                           0.81     30000\n",
      "   macro avg       0.77      0.61      0.63     30000\n",
      "weighted avg       0.80      0.81      0.78     30000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "pred = gs_model.best_estimator_.predict(pca_scaled.transform(X_scaled))\n",
    "print(classification_report(y, pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### With all data this seems to work.... however, we still have the feeling that this is too good to be true and that there might be a chance that the model is overfitting."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Random Forest Classifier\n",
    "## ==================================================================="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "forest_param_grid = {\n",
    "  'criterion': ['gini'],\n",
    "  'max_depth': [6],\n",
    "  'n_estimators': [150],  \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 14.998s:\n",
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=6, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=2,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=150,\n",
      "                       n_jobs=None, oob_score=False, random_state=None,\n",
      "                       verbose=0, warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      1.00      0.87      8134\n",
      "           1       0.50      0.01      0.01      2366\n",
      "\n",
      "    accuracy                           0.77     10500\n",
      "   macro avg       0.64      0.50      0.44     10500\n",
      "weighted avg       0.71      0.77      0.68     10500\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with RandomForest to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(RandomForestClassifier(), forest_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 15.398s:\n",
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=6, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=2,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=150,\n",
      "                       n_jobs=None, oob_score=False, random_state=None,\n",
      "                       verbose=0, warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      1.00      0.87      8134\n",
      "           1       0.00      0.00      0.00      2366\n",
      "\n",
      "    accuracy                           0.77     10500\n",
      "   macro avg       0.39      0.50      0.44     10500\n",
      "weighted avg       0.60      0.77      0.68     10500\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/alemadlei/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with RandomForest to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(RandomForestClassifier(), forest_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_norm_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_norm_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 15.716s:\n",
      "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
      "                       max_depth=6, max_features='auto', max_leaf_nodes=None,\n",
      "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
      "                       min_samples_leaf=1, min_samples_split=2,\n",
      "                       min_weight_fraction_leaf=0.0, n_estimators=150,\n",
      "                       n_jobs=None, oob_score=False, random_state=None,\n",
      "                       verbose=0, warm_start=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.96      0.90      8134\n",
      "           1       0.75      0.44      0.55      2366\n",
      "\n",
      "    accuracy                           0.84     10500\n",
      "   macro avg       0.80      0.70      0.73     10500\n",
      "weighted avg       0.83      0.84      0.82     10500\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with RandomForest to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(RandomForestClassifier(), forest_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_scaled_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_scaled_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Support Vector Machines Evaluation\n",
    "## ==================================================================="
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We already have all of our data transformed and ready to use.\n",
    "# Build the parameters\n",
    "svm_param_grid = {\n",
    "  'kernel': ['rbf'],\n",
    "  'gamma': [0.0001, 0.0005, 0.001, 0.005],\n",
    "  'class_weight': ['balanced']\n",
    "}\n",
    "\n",
    "svm_param_grid = {\n",
    "  'kernel': ['rbf'],\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 42.674s:\n",
      "SVC(C=1.0, cache_size=2048, class_weight=None, coef0=0.0,\n",
      "    decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',\n",
      "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "    tol=0.001, verbose=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      1.00      0.87      8134\n",
      "           1       0.00      0.00      0.00      2366\n",
      "\n",
      "    accuracy                           0.77     10500\n",
      "   macro avg       0.39      0.50      0.44     10500\n",
      "weighted avg       0.60      0.77      0.68     10500\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/alemadlei/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with SVM to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(SVC(cache_size=2048, gamma= 'scale'), svm_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 36.379s:\n",
      "SVC(C=1.0, cache_size=2048, class_weight=None, coef0=0.0,\n",
      "    decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',\n",
      "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "    tol=0.001, verbose=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      1.00      0.87      8134\n",
      "           1       0.00      0.00      0.00      2366\n",
      "\n",
      "    accuracy                           0.77     10500\n",
      "   macro avg       0.39      0.50      0.44     10500\n",
      "weighted avg       0.60      0.77      0.68     10500\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/alemadlei/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with SVM to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(SVC(cache_size=2048, gamma= 'scale'), svm_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_norm_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_norm_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best estimator found in 78.522s:\n",
      "SVC(C=1.0, cache_size=2048, class_weight=None, coef0=0.0,\n",
      "    decision_function_shape='ovr', degree=3, gamma='scale', kernel='rbf',\n",
      "    max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
      "    tol=0.001, verbose=False)\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.96      0.89      8134\n",
      "           1       0.71      0.35      0.47      2366\n",
      "\n",
      "    accuracy                           0.82     10500\n",
      "   macro avg       0.77      0.66      0.68     10500\n",
      "weighted avg       0.81      0.82      0.80     10500\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Uses grid search with SVM to obtain the best estimator.\n",
    "t0 = time()\n",
    "gs_model = GridSearchCV(SVC(cache_size=2048, gamma= 'scale'), svm_param_grid, cv=10, iid=False)\n",
    "gs_model = gs_model.fit(X_scaled_train, y_train)\n",
    "\n",
    "# Dislay messages.\n",
    "print('Best estimator found in %0.3fs:' % (time() - t0))\n",
    "print(gs_model.best_estimator_)\n",
    "\n",
    "# Using test set.\n",
    "y_pred = gs_model.predict(X_scaled_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "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": 2
}