Logistic_Regression_Keras.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "__author__ = \"Neelabh Pant\"\n",
    "## Date: 06/14/2017"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project Description\n",
    "\n",
    "A Lot of people leaving the company recently and we would like to understand why that's happening. We have historical data on employees and in this project we would like to build a model that is able to predict which employee will leave next. We have built a model that is better than random guessing. Here in this project we also prefer false negatives than false positives. Fields in the dataset include:\n",
    "* Employee satisfaction level\n",
    "* Last evaluation\n",
    "* Number of projects\n",
    "* Average monthly hours\n",
    "* Time spent at the company\n",
    "* Whether they have had a work accident\n",
    "* Whether they have had a promotion in the last 5 years\n",
    "* Department\n",
    "* Salary\n",
    "* Whether the employee has left\n",
    "\n",
    "The goal of this project is to predict the binary outcome variable \"left\" using the rest of the data. Since the outcome is binary, this is a classification problem.\n",
    "\n",
    "This dataset comes from (https://www.kaggle.com/ludobenistant/hr-analytics/) and is released under CC BY-SA 4.0 License."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from __future__ import division"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils import to_categorical\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.optimizers import SGD, Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('/Users/neelabhpant/Documents/Deep Learning/zero_to_deep_learning_udemy/data/HR_comma_sep.csv')\n",
    "df_toplot = df.copy()\n",
    "df_tocopy=df_toplot.drop(['salary', 'sales'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEFCAYAAAD5bXAgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x124be91d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAKYAAAB1CAYAAADEIpr2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noipJipmoSpJiJqqSpJiJqiQpZqIq+X+jTKaDJ+HIYAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x125f500f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAKEAAAB1CAYAAAAm/oGPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1200a5d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAALQAAAB6CAYAAAAWL3n2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQidqiqTQiZri/wHQ9u1vDr+RHQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1241b8240>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAK8AAAB6CAYAAADJYQPpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x125f54400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAKQAAAB6CAYAAAAxgfgeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10dbdbb38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAALQAAAB6CAYAAAAWL3n2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x125f1e208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAALYAAAB6CAYAAAAS2qnLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUbEjmSQqdiSTRMWOZJKo2JFM8n+aHxpdztFB0gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1268c42b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15, 10))\n",
    "for i, feature in enumerate(df_tocopy.columns):\n",
    "    plt.subplot(3, 3, i+1)\n",
    "    df[feature].plot(kind='hist', title=feature)\n",
    "    plt.xlabel(feature)\n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cleaning data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df['monthly_hours_norm'] = df['average_montly_hours']/df['average_montly_hours'].max()\n",
    "y = df['left'].as_matrix()\n",
    "y_1 = to_categorical(df['left'])\n",
    "Work_accident_categorical = to_categorical(df['Work_accident'])\n",
    "promotion_last_5years_categorical = to_categorical(df['promotion_last_5years'])\n",
    "\n",
    "positions=np.zeros((df.sales.shape[0], 1))\n",
    "for idx,name in enumerate(df.sales):\n",
    "    if name=='sales':\n",
    "        positions[idx,0]=0\n",
    "    elif name=='accounting':\n",
    "        positions[idx,0]=1\n",
    "    elif name=='hr':\n",
    "        positions[idx,0]=2\n",
    "    elif name=='technical':\n",
    "        positions[idx,0]=3\n",
    "    elif name=='support':\n",
    "        positions[idx,0]=4\n",
    "    elif name=='management':\n",
    "        positions[idx,0]=5\n",
    "    elif name=='IT':\n",
    "        positions[idx,0]=6\n",
    "    elif name=='product_mng':\n",
    "        positions[idx,0]=7\n",
    "    elif name=='marketing':\n",
    "        positions[idx,0]=8\n",
    "    else:\n",
    "        positions[idx,0]=9\n",
    "        \n",
    "#df_positions = pd.DataFrame(data=positions, columns=['positions'])\n",
    "categorical_positions = to_categorical(positions)\n",
    "\n",
    "monthly_salary=np.zeros((df.salary.shape[0], 1))\n",
    "for idx,name in enumerate(df.salary):\n",
    "    if name=='low':\n",
    "        monthly_salary[idx,0]=0\n",
    "    elif name=='medium':\n",
    "        monthly_salary[idx,0]=1\n",
    "    else:\n",
    "        monthly_salary[idx,0]=2\n",
    "\n",
    "categorical_salary = to_categorical(monthly_salary)\n",
    "\n",
    "df = df[['satisfaction_level', 'last_evaluation', 'number_project',\n",
    "         'time_spend_company', 'monthly_hours_norm']]\n",
    "\n",
    "df['No Accident'] = Work_accident_categorical[:,0]\n",
    "df['Accident'] = Work_accident_categorical[:,1]\n",
    "df['No promotion'] = promotion_last_5years_categorical[:,0]\n",
    "df['Promotion'] = promotion_last_5years_categorical[:,1]\n",
    "df['sales'] = categorical_positions[:,0]\n",
    "df['accounting'] = categorical_positions[:,1]\n",
    "df['hr'] = categorical_positions[:,2]\n",
    "df['technical'] = categorical_positions[:,3]\n",
    "df['support'] = categorical_positions[:,4]\n",
    "df['management'] = categorical_positions[:,5]\n",
    "df['IT'] = categorical_positions[:,6]\n",
    "df['product_mng'] = categorical_positions[:,7]\n",
    "df['marketing'] = categorical_positions[:,8]\n",
    "df['random'] = categorical_positions[:,9]\n",
    "df['low_salary'] = categorical_salary[:,0]\n",
    "df['medium_salary'] = categorical_salary[:,1]\n",
    "df['high_salary'] = categorical_salary[:,2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = df[['satisfaction_level', 'last_evaluation', 'number_project',\n",
    "       'time_spend_company', 'monthly_hours_norm', 'No Accident', 'Accident',\n",
    "       'No promotion', 'Promotion', 'sales', 'accounting', 'hr', 'technical',\n",
    "       'support', 'management', 'IT', 'product_mng', 'marketing', 'random',\n",
    "       'low_salary', 'medium_salary', 'high_salary']].as_matrix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dividing data into training and testing sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Intializing and Training Logistic Regression model using Keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(10, input_shape=(X_train.shape[1],), activation='sigmoid'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.compile(optimizer=Adam(lr=0.1), loss='binary_crossentropy', metrics=['accuracy'])\n",
    "model.fit(X_train, y_train, epochs=40, verbose=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting outcomes using trained model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train_pred = model.predict(X_train)\n",
    "y_train_class_pred = y_train_pred > 0.5\n",
    "\n",
    "y_test_pred = model.predict(X_test)\n",
    "y_test_class_pred = y_test_pred > 0.5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Accuracy check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The training accuracy score is 0.956\n",
      "The testing accuracy score is 0.950\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "print(\"The training accuracy score is {:0.3f}\".format(accuracy_score(y_train, y_train_class_pred)))\n",
    "print(\"The testing accuracy score is {:0.3f}\".format(accuracy_score(y_test, y_test_class_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Data Confusion Matrix :\n",
      "\n",
      "           Predicted Left  Predicted Not Left\n",
      "Left                8901                 253\n",
      "Not Left             269                2576\n",
      "\n",
      "\n",
      "Testing Data Confusion Matrix :\n",
      "\n",
      "           Predicted Left  Predicted Not Left\n",
      "Left                2203                  71\n",
      "Not Left              79                 647\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "def pretty_confusion_matrix(y_true, y_pred, labels=[\"False\", \"True\"]):\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    pred_labels = [\"Predicted \"+ l for l in labels]\n",
    "    df = pd.DataFrame(cm, index=labels, columns=pred_labels)\n",
    "    return df\n",
    "\n",
    "print(\"Training Data Confusion Matrix :\\n\\n\",pretty_confusion_matrix(y_train, y_train_class_pred, ['Left', \"Not Left\"]))\n",
    "print(\"\\n\")\n",
    "print(\"Testing Data Confusion Matrix :\\n\\n\",pretty_confusion_matrix(y_test, y_test_class_pred, ['Left', \"Not Left\"]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Precision, Recall and R-2 score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The R2 score on the Train set is:\t0.792\n",
      "The R2 score on the Test set is:\t0.762\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import r2_score\n",
    "print(\"The R2 score on the Train set is:\\t{:0.3f}\".format(r2_score(y_train, y_train_pred)))\n",
    "print(\"The R2 score on the Test set is:\\t{:0.3f}\".format(r2_score(y_test, y_test_pred)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# K-fold cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The cross validation accuracy is 0.9422 ± 0.0188\n"
     ]
    }
   ],
   "source": [
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.model_selection import cross_val_score, KFold\n",
    "\n",
    "def build_logistic_regression_model():\n",
    "    model = Sequential()\n",
    "    model.add(Dense(10, input_shape=(X_train.shape[1],), activation='sigmoid'))\n",
    "    model.add(Dense(1, activation='sigmoid'))\n",
    "    model.compile(Adam(lr=0.1),\n",
    "                  'binary_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "    return model\n",
    "\n",
    "model = KerasClassifier(build_fn=build_logistic_regression_model,\n",
    "                        epochs=25,\n",
    "                        verbose=0)\n",
    "cv = KFold(5, shuffle=True)\n",
    "scores = cross_val_score(model, X_train, y_train, cv=cv)\n",
    "print(\"The cross validation accuracy is {:0.4f} ± {:0.4f}\".format(scores.mean(), scores.std()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "The 5-Fold Cross Validation accuracy was almost equal to the Accuracy . \n",
    "Hence we can assume the model is doing good even when the data is shuffled and presented training and testing data in K-Folds\n",
    "\n",
    "-Neelabh Pant"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}