09 Neural Networks (Classification).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img align=\"center\" src=\"http://sydney.edu.au/images/content/about/logo-mono.jpg\">\n",
    "<h1 align=\"center\" style=\"margin-top:10px\">Statistical Learning with Python</h1>\n",
    "<h2 align=\"center\" style=\"margin-top:20px\">Lecture 9: Neural Networks (Classification)</h2>\n",
    "<br>\n",
    "\n",
    "<a href=\"#1.-HR-Analytics-Data\">HR Analytics Data</a> <br>\n",
    "<a href=\"#2.-Feature-Engineering\">Feature Engineering</a> <br>\n",
    "<a href=\"#3.-Neural-Network\">Neural Network</a> <br>\n",
    "<a href=\"#4.-Dropout\">Dropout</a> <br>\n",
    "<a href=\"#5.-Other-Methods\">Other Methods</a> <br>\n",
    "<a href=\"#6.-Model-Evaluation\">Model Evaluation</a> <br>\n",
    "\n",
    "This notebook relies on the following libraries and settings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Packages\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore') \n",
    "from IPython.display import clear_output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot settings\n",
    "sns.set_context('notebook') # optimises figures for notebook display\n",
    "sns.set_style('ticks') # set default plot style\n",
    "crayon = ['#4E79A7','#F28E2C','#E15759','#76B7B2','#59A14F', \n",
    "          '#EDC949','#AF7AA1','#FF9DA7','#9C755F','#BAB0AB']\n",
    "sns.set_palette(crayon) # set custom color scheme\n",
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (9, 6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import RandomizedSearchCV, GridSearchCV\n",
    "from sklearn.metrics import accuracy_score, recall_score, precision_score, roc_auc_score\n",
    "from sklearn.metrics import confusion_matrix, log_loss, average_precision_score\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression, LogisticRegressionCV\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import  RandomForestClassifier\n",
    "import lightgbm as lgb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. HR Analytics Data\n",
    "\n",
    "We use the [Human Resources Analytics](https://www.kaggle.com/ludobenistant/hr-analytics) data available from [Kaggle Datasets](https://www.kaggle.com/datasets). Our objective is to predict To predict which employees will leave the company."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>satisfaction_level</th>\n",
       "      <th>last_evaluation</th>\n",
       "      <th>number_project</th>\n",
       "      <th>average_montly_hours</th>\n",
       "      <th>time_spend_company</th>\n",
       "      <th>Work_accident</th>\n",
       "      <th>left</th>\n",
       "      <th>promotion_last_5years</th>\n",
       "      <th>role</th>\n",
       "      <th>salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.38</td>\n",
       "      <td>0.53</td>\n",
       "      <td>2</td>\n",
       "      <td>157</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sales</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.80</td>\n",
       "      <td>0.86</td>\n",
       "      <td>5</td>\n",
       "      <td>262</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sales</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.11</td>\n",
       "      <td>0.88</td>\n",
       "      <td>7</td>\n",
       "      <td>272</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sales</td>\n",
       "      <td>medium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.72</td>\n",
       "      <td>0.87</td>\n",
       "      <td>5</td>\n",
       "      <td>223</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sales</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.37</td>\n",
       "      <td>0.52</td>\n",
       "      <td>2</td>\n",
       "      <td>159</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>sales</td>\n",
       "      <td>low</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   satisfaction_level  last_evaluation  number_project  average_montly_hours  \\\n",
       "0                0.38             0.53               2                   157   \n",
       "1                0.80             0.86               5                   262   \n",
       "2                0.11             0.88               7                   272   \n",
       "3                0.72             0.87               5                   223   \n",
       "4                0.37             0.52               2                   159   \n",
       "\n",
       "   time_spend_company  Work_accident  left  promotion_last_5years   role  \\\n",
       "0                   3              0     1                      0  sales   \n",
       "1                   6              0     1                      0  sales   \n",
       "2                   4              0     1                      0  sales   \n",
       "3                   5              0     1                      0  sales   \n",
       "4                   3              0     1                      0  sales   \n",
       "\n",
       "   salary  \n",
       "0     low  \n",
       "1  medium  \n",
       "2  medium  \n",
       "3     low  \n",
       "4     low  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('Data/HR.csv')\n",
    "data = data.rename(columns = {'sales' : 'role'})\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(14999, 10)\n"
     ]
    }
   ],
   "source": [
    "print(data.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "response='left'\n",
    "index_train, index_test  = train_test_split(np.array(data.index), stratify=data[response], \n",
    "                                            train_size=0.5, random_state=5)\n",
    "\n",
    "y_train = data.loc[index_train, response].to_numpy()\n",
    "y_test = data.loc[index_test, response].to_numpy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Feature Engineering\n",
    "\n",
    "Before training the models, we need to convert the categorical variables into binary variables. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label in ['role', 'salary']:\n",
    "    dummies = pd.get_dummies(data[label], drop_first=True)\n",
    "    data = data.join(dummies)\n",
    "    data = data.drop(label, axis= 1)\n",
    "    \n",
    "predictors = list(data.columns)\n",
    "predictors.remove(response)\n",
    "\n",
    "train = data.loc[index_train,:].copy()\n",
    "test =  data.loc[index_test,:].copy()\n",
    "\n",
    "X_train = data.loc[index_train, predictors].to_numpy()\n",
    "X_test = data.loc[index_test, predictors].to_numpy()\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaler.fit(X_train)\n",
    "\n",
    "X_train = scaler.transform(X_train)\n",
    "X_test = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(7499, 18)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Neural Network\n",
    "\n",
    "**3.1 Specifying the model** \n",
    "\n",
    "We specify a multilayer perceptron with two hidden layers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.engine.sequential.Sequential at 0x1791b75dac8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "mlp = Sequential()\n",
    "mlp.add(Dense(128, input_dim=X_train.shape[1], activation='relu'))\n",
    "mlp.add(Dense(64, activation='relu'))\n",
    "mlp.add(Dense(1, activation='sigmoid'))\n",
    "mlp.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])\n",
    "clear_output()\n",
    "mlp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**3.2 Creating a Validation Set** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_val = int(np.floor(0.2*len(y_train)))\n",
    "\n",
    "X_train_partial = X_train[:-n_val,:]\n",
    "y_train_partial = y_train[:-n_val]\n",
    "X_val =  X_train[-n_val:,:]\n",
    "y_val =  y_train[-n_val:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**3.3 Training the Neural Network** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\marce\\Anaconda3\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n"
     ]
    }
   ],
   "source": [
    "fitted = mlp.fit(X_train_partial, y_train_partial, epochs=80, validation_data=(X_val, y_val), verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['val_loss', 'val_acc', 'loss', 'acc'])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history = fitted.history\n",
    "history.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax= plt.subplots()\n",
    "ax.plot(history['loss'], label='Training loss')\n",
    "ax.plot(history['val_loss'], label='Validation loss')\n",
    "ax.set_xlabel('Epoch')\n",
    "ax.set_ylabel('Loss')\n",
    "ax.legend(frameon=False)\n",
    "sns.despine()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax= plt.subplots()\n",
    "ax.plot(1-np.array(history['acc']), label='Training loss')\n",
    "ax.plot(1-np.array(history['val_acc']), label='Validation loss')\n",
    "ax.set_xlabel('Epoch')\n",
    "ax.set_ylabel('Loss')\n",
    "ax.set_title('Error Rate')\n",
    "ax.legend(frameon=False)\n",
    "sns.despine()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "7499/7499 [==============================] - 1s 82us/step - loss: 0.2981 - acc: 0.8820\n",
      "Epoch 2/20\n",
      "7499/7499 [==============================] - 0s 41us/step - loss: 0.1809 - acc: 0.9459\n",
      "Epoch 3/20\n",
      "7499/7499 [==============================] - 0s 39us/step - loss: 0.1629 - acc: 0.9529\n",
      "Epoch 4/20\n",
      "7499/7499 [==============================] - 0s 41us/step - loss: 0.1530 - acc: 0.9545\n",
      "Epoch 5/20\n",
      "7499/7499 [==============================] - 0s 39us/step - loss: 0.1457 - acc: 0.9563\n",
      "Epoch 6/20\n",
      "7499/7499 [==============================] - 0s 43us/step - loss: 0.1399 - acc: 0.9587\n",
      "Epoch 7/20\n",
      "7499/7499 [==============================] - 0s 39us/step - loss: 0.1346 - acc: 0.9601\n",
      "Epoch 8/20\n",
      "7499/7499 [==============================] - 0s 38us/step - loss: 0.1280 - acc: 0.9621\n",
      "Epoch 9/20\n",
      "7499/7499 [==============================] - 0s 39us/step - loss: 0.1251 - acc: 0.9636\n",
      "Epoch 10/20\n",
      "7499/7499 [==============================] - 0s 40us/step - loss: 0.1193 - acc: 0.9641\n",
      "Epoch 11/20\n",
      "7499/7499 [==============================] - 0s 38us/step - loss: 0.1150 - acc: 0.9656\n",
      "Epoch 12/20\n",
      "7499/7499 [==============================] - 0s 40us/step - loss: 0.1108 - acc: 0.9669\n",
      "Epoch 13/20\n",
      "7499/7499 [==============================] - 0s 37us/step - loss: 0.1093 - acc: 0.9680\n",
      "Epoch 14/20\n",
      "7499/7499 [==============================] - 0s 46us/step - loss: 0.1051 - acc: 0.9679\n",
      "Epoch 15/20\n",
      "7499/7499 [==============================] - 0s 40us/step - loss: 0.1008 - acc: 0.9695\n",
      "Epoch 16/20\n",
      "7499/7499 [==============================] - 0s 42us/step - loss: 0.0980 - acc: 0.9703\n",
      "Epoch 17/20\n",
      "7499/7499 [==============================] - 0s 42us/step - loss: 0.0941 - acc: 0.9724\n",
      "Epoch 18/20\n",
      "7499/7499 [==============================] - 0s 42us/step - loss: 0.0911 - acc: 0.9716\n",
      "Epoch 19/20\n",
      "7499/7499 [==============================] - 0s 41us/step - loss: 0.0900 - acc: 0.9732\n",
      "Epoch 20/20\n",
      "7499/7499 [==============================] - 0s 38us/step - loss: 0.0871 - acc: 0.9735\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x17923e51a58>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mlp = Sequential()\n",
    "mlp.add(Dense(128, input_dim=X_train.shape[1], activation='relu'))\n",
    "mlp.add(Dense(64, activation='relu'))\n",
    "mlp.add(Dense(1, activation='sigmoid'))\n",
    "mlp.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])\n",
    "mlp.fit(X_train, y_train, epochs=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\marce\\Anaconda3\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x179262b6c50>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.layers import Dropout\n",
    "\n",
    "mlp_dropout = Sequential()\n",
    "mlp_dropout.add(Dense(128, input_dim=X_train.shape[1], activation='relu'))\n",
    "mlp_dropout.add(Dropout(0.5))\n",
    "mlp_dropout.add(Dense(64, activation='relu'))\n",
    "mlp_dropout.add(Dense(1, activation='sigmoid'))\n",
    "mlp_dropout.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])\n",
    "mlp_dropout.fit(X_train, y_train, epochs=200, verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Other Methods\n",
    "\n",
    "**5.1 Logistic Regression**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=inf, 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='lbfgs',\n",
       "          tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "logit = LogisticRegression(C=np.inf, solver='lbfgs')\n",
    "logit.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**5.2 Random Forest**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameters found by randomised search: {'min_samples_leaf': 5, 'max_features': 5} \n",
      "\n",
      "Wall time: 26.7 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "model = RandomForestClassifier(criterion = 'entropy',  n_estimators=100)\n",
    "\n",
    "tuning_parameters = {\n",
    "    'min_samples_leaf': [1, 5, 10, 20, 50],\n",
    "    'max_features': np.arange(1, X_train.shape[1]),\n",
    "}\n",
    "\n",
    "rf_search = RandomizedSearchCV(model, tuning_parameters, cv = 5, n_iter= 32, scoring='neg_log_loss', n_jobs=4)\n",
    "rf_search.fit(X_train, y_train)\n",
    "\n",
    "rf = rf_search.best_estimator_\n",
    "\n",
    "print('Best parameters found by randomised search:', rf_search.best_params_, '\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='entropy',\n",
       "            max_depth=None, max_features=5, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=5, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=1000, n_jobs=None,\n",
       "            oob_score=False, random_state=1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf.n_estimators = 1000\n",
    "rf.random_state = 1\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**5.3 Gradient Boosting**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost import XGBClassifier\n",
    "from skopt.space import Real, Categorical, Integer\n",
    "from skopt import BayesSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best parameters found by Bayesian estimation: {'gamma': 3.589181775873712e-06, 'learning_rate': 0.005, 'max_depth': 8, 'n_estimators': 1551, 'reg_lambda': 1.9870009752320832e-12, 'subsample': 1.0} \n",
      "\n",
      "Wall time: 10min 46s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "model = XGBClassifier()\n",
    "\n",
    "space = {\n",
    "    'reg_lambda':  (1e-15, 1000, 'log-uniform'),\n",
    "    'gamma':  (1e-15, 1000, 'log-uniform'),\n",
    "    'learning_rate': Real(0.005, 0.15),\n",
    "    'n_estimators' : Integer(100, 2000),\n",
    "    'max_depth' : Integer(2, 10),\n",
    "    'subsample' : Real(0.5, 1.0),\n",
    "}\n",
    "\n",
    "np.random.seed(87)\n",
    "xgb_search =  BayesSearchCV(model, space, cv = 5,  n_iter=32, scoring = 'neg_log_loss', n_jobs=4)\n",
    "xgb_search.fit(X_train, y_train)\n",
    "\n",
    "xgb = xgb_search.best_estimator_\n",
    "\n",
    "print('Best parameters found by Bayesian estimation:', xgb_search.best_params_, '\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Model Evaluation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "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>Error rate</th>\n",
       "      <th>Sensitivity</th>\n",
       "      <th>Specificity</th>\n",
       "      <th>AUC</th>\n",
       "      <th>Precision</th>\n",
       "      <th>Average Precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Logistic regression</th>\n",
       "      <td>0.211</td>\n",
       "      <td>0.342</td>\n",
       "      <td>0.929</td>\n",
       "      <td>0.827</td>\n",
       "      <td>0.601</td>\n",
       "      <td>0.533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>0.024</td>\n",
       "      <td>0.911</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.992</td>\n",
       "      <td>0.985</td>\n",
       "      <td>0.984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XGBoost</th>\n",
       "      <td>0.017</td>\n",
       "      <td>0.943</td>\n",
       "      <td>0.995</td>\n",
       "      <td>0.994</td>\n",
       "      <td>0.984</td>\n",
       "      <td>0.987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Neural Net</th>\n",
       "      <td>0.038</td>\n",
       "      <td>0.927</td>\n",
       "      <td>0.973</td>\n",
       "      <td>0.983</td>\n",
       "      <td>0.915</td>\n",
       "      <td>0.954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Neural Net (dropout)</th>\n",
       "      <td>0.031</td>\n",
       "      <td>0.901</td>\n",
       "      <td>0.990</td>\n",
       "      <td>0.978</td>\n",
       "      <td>0.967</td>\n",
       "      <td>0.960</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      Error rate  Sensitivity  Specificity    AUC  Precision  \\\n",
       "Logistic regression        0.211        0.342        0.929  0.827      0.601   \n",
       "Random Forest              0.024        0.911        0.996  0.992      0.985   \n",
       "XGBoost                    0.017        0.943        0.995  0.994      0.984   \n",
       "Neural Net                 0.038        0.927        0.973  0.983      0.915   \n",
       "Neural Net (dropout)       0.031        0.901        0.990  0.978      0.967   \n",
       "\n",
       "                      Average Precision  \n",
       "Logistic regression               0.533  \n",
       "Random Forest                     0.984  \n",
       "XGBoost                           0.987  \n",
       "Neural Net                        0.954  \n",
       "Neural Net (dropout)              0.960  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns=['Error rate', 'Sensitivity', 'Specificity', 'AUC', 'Precision', 'Average Precision']\n",
    "rows=['Logistic regression', 'Random Forest', 'XGBoost', 'Neural Net', 'Neural Net (dropout)']\n",
    "results=pd.DataFrame(0.0, columns=columns, index=rows) \n",
    "\n",
    "methods=[logit, rf, xgb, mlp, mlp_dropout]\n",
    "\n",
    "for i, method in enumerate(methods):\n",
    "    \n",
    "    y_pred = method.predict(X_test)\n",
    "    if (method == mlp) | (method == mlp_dropout):\n",
    "        y_prob = y_pred\n",
    "        y_pred = (y_pred>0.5).astype(int)\n",
    "    else:\n",
    "        y_prob = method.predict_proba(X_test)[:,1]\n",
    "        \n",
    "\n",
    "\n",
    "    confusion  = confusion_matrix(y_test, y_pred)\n",
    "    results.iloc[i,0]=  1 - accuracy_score(y_test, y_pred)\n",
    "    results.iloc[i,1]=  confusion[1,1]/np.sum(confusion[1,:])\n",
    "    results.iloc[i,2]=  confusion[0,0]/np.sum(confusion[0,:])\n",
    "    results.iloc[i,4]=  precision_score(y_test, y_pred)\n",
    "    results.iloc[i,3] = roc_auc_score(y_test, y_prob)\n",
    "    results.iloc[i,5]=  average_precision_score(y_test, y_prob)\n",
    "\n",
    "results.round(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Additional Code\n",
    "\n",
    "The two cells below format the notebook for display online. Please omit them from your work."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>\n",
       "@import url('https://fonts.googleapis.com/css?family=Source+Sans+Pro|Open+Sans:800&display=swap');\n",
       "</style>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%html\n",
    "<style>\n",
    "@import url('https://fonts.googleapis.com/css?family=Source+Sans+Pro|Open+Sans:800&display=swap');\n",
    "</style>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.CodeMirror pre {font-family: monospace; font-size: 12pt; line-height: 22.4px;}\n",
       "\n",
       "pre {font-family: monospace; font-size: 12pt; line-height: 22.4px;}\n",
       "\n",
       ".output_area pre {font-family: monospace; font-size: 12pt; line-height: 22.4px;}\n",
       "\n",
       ".rendered_html table {font-size: 10.5pt;}\n",
       "\n",
       "h1 {\n",
       "  font-family: 'Open Sans' !important;\n",
       "  color: #222;\n",
       "  font-size: 2.5em !important;\n",
       "  padding-bottom: 0px;  \n",
       "  margin-top: 14px;\n",
       "  margin-bottom: 6px;\n",
       "  font-weight: 800 !important;\n",
       "}\n",
       "\n",
       "\n",
       "h2 {\n",
       "  font-family: 'Open Sans' !important;\n",
       "  font-size: 1.6em !important;\n",
       "  font-weight: 800 !important;\n",
       "  color: #222;\n",
       "}\n",
       "\n",
       "\n",
       "h3 {\n",
       "  font-family: 'Open Sans' !important;\n",
       "  font-size: 25px !important;\n",
       "  color: #222 !important;\n",
       "  margin-bottom: 30px !important;\n",
       "  margin-top: 25px !important;\n",
       "}\n",
       "\n",
       "\n",
       "p {\n",
       "  font-family: 'Source Sans Pro', sans-serif;\n",
       "  color: #222;\n",
       "  font-size: 19px;\n",
       "  margin-bottom: 10px;\n",
       "  margin-top:0;\n",
       "  line-height: 1.4;\n",
       "}\n",
       "\n",
       "li {\n",
       "  font-family: 'Source Sans Pro', sans-serif;\n",
       "  color: #222;\n",
       "  font-size: 19px;\n",
       "  line-height: 1.4;\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "/*code {\n",
       "     padding: 2px 4px !important; \n",
       "     font-size: 90% !important;\n",
       "     color: #c7254e !important; \n",
       "     background-color: #f9f2f4 !important; \n",
       "     border-radius: 2px !important;\n",
       "     text-decoration: none !important;\n",
       "      font-size: 90% !important;\n",
       "}\n",
       "*/\n",
       "\n",
       "\n",
       "code {\n",
       "      font-size: 90% !important;\n",
       "/*      border-style: solid !important;\n",
       "      border-width: 1px !important;\n",
       "      border-color :#cccccc !important;*/\n",
       "}\n",
       "\n",
       "TT {\n",
       "      font-family: monospace;\n",
       "      font-size: 95% !important;\n",
       "}\n",
       "\n",
       "\n",
       "\n",
       "a.neutral {\n",
       "  text-decoration: none !important;\n",
       "  border-style: dotted !important;\n",
       "  border-width: 1px !important;\n",
       "  border-color :#cccccc !important;\n",
       "\n",
       "}\n",
       "\n",
       "</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.core.display import HTML\n",
    "style = open('jstyle.css', \"r\").read()\n",
    "HTML('<style>'+ style +'</style>')"
   ]
  }
 ],
 "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
}