xgboost_visualizer_demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:34.448491Z",
     "start_time": "2018-12-07T18:03:33.698447Z"
    }
   },
   "outputs": [],
   "source": [
    "# import\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import xgboost as xgb\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from xgboost_visualizer import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:34.487782Z",
     "start_time": "2018-12-07T18:03:34.450800Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[13:03:34] 12000x9 matrix with 108000 entries loaded from data/train.libsvm\n",
      "[13:03:34] 2999x9 matrix with 26991 entries loaded from data/test.libsvm\n"
     ]
    }
   ],
   "source": [
    "# load data\n",
    "dtrain = xgb.DMatrix(\"data/train.libsvm\")\n",
    "dtest = xgb.DMatrix(\"data/test.libsvm\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:34.494423Z",
     "start_time": "2018-12-07T18:03:34.490076Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train label distribution: 0.23783333333333334\n",
      "Train label distribution: 0.23907969323107703\n"
     ]
    }
   ],
   "source": [
    "# train data label distrubtion\n",
    "print(\"Train label distribution:\", dtrain.get_label().sum() / dtrain.get_label().size)\n",
    "\n",
    "# test data label distrubtion\n",
    "print(\"Train label distribution:\", dtest.get_label().sum() / dtest.get_label().size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:34.500733Z",
     "start_time": "2018-12-07T18:03:34.497472Z"
    }
   },
   "outputs": [],
   "source": [
    "# define parameters\n",
    "params = {\n",
    "    \"objective\": \"binary:logistic\",\n",
    "    \"silent\": 1,\n",
    "    \"eval_metric\": \"auc\"\n",
    "}\n",
    "\n",
    "watchlist = [(dtrain, 'train'), (dtest, 'eval')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:35.257671Z",
     "start_time": "2018-12-07T18:03:34.502835Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-auc:0.886794\teval-auc:0.881745\n",
      "[1]\ttrain-auc:0.898009\teval-auc:0.887637\n",
      "[2]\ttrain-auc:0.908018\teval-auc:0.895209\n",
      "[3]\ttrain-auc:0.913157\teval-auc:0.897542\n",
      "[4]\ttrain-auc:0.916913\teval-auc:0.897553\n",
      "[5]\ttrain-auc:0.920597\teval-auc:0.901308\n",
      "[6]\ttrain-auc:0.924696\teval-auc:0.904203\n",
      "[7]\ttrain-auc:0.926771\teval-auc:0.90488\n",
      "[8]\ttrain-auc:0.928893\teval-auc:0.905824\n",
      "[9]\ttrain-auc:0.930563\teval-auc:0.90824\n",
      "[10]\ttrain-auc:0.932033\teval-auc:0.909653\n",
      "[11]\ttrain-auc:0.933475\teval-auc:0.910952\n",
      "[12]\ttrain-auc:0.934853\teval-auc:0.91125\n",
      "[13]\ttrain-auc:0.936429\teval-auc:0.912229\n",
      "[14]\ttrain-auc:0.936603\teval-auc:0.913077\n",
      "[15]\ttrain-auc:0.937018\teval-auc:0.913533\n",
      "[16]\ttrain-auc:0.937963\teval-auc:0.913971\n",
      "[17]\ttrain-auc:0.93907\teval-auc:0.916046\n",
      "[18]\ttrain-auc:0.939861\teval-auc:0.916191\n",
      "[19]\ttrain-auc:0.940736\teval-auc:0.916247\n",
      "[20]\ttrain-auc:0.941288\teval-auc:0.917188\n",
      "[21]\ttrain-auc:0.942817\teval-auc:0.917256\n",
      "[22]\ttrain-auc:0.943014\teval-auc:0.917674\n",
      "[23]\ttrain-auc:0.943805\teval-auc:0.917472\n",
      "[24]\ttrain-auc:0.944767\teval-auc:0.917824\n",
      "[25]\ttrain-auc:0.945321\teval-auc:0.918311\n",
      "[26]\ttrain-auc:0.945387\teval-auc:0.918572\n",
      "[27]\ttrain-auc:0.946157\teval-auc:0.918083\n",
      "[28]\ttrain-auc:0.947588\teval-auc:0.918648\n",
      "[29]\ttrain-auc:0.947797\teval-auc:0.919223\n",
      "[30]\ttrain-auc:0.948544\teval-auc:0.919216\n",
      "[31]\ttrain-auc:0.948965\teval-auc:0.918958\n",
      "[32]\ttrain-auc:0.949288\teval-auc:0.918936\n",
      "[33]\ttrain-auc:0.949913\teval-auc:0.919177\n",
      "[34]\ttrain-auc:0.950538\teval-auc:0.918852\n",
      "[35]\ttrain-auc:0.950745\teval-auc:0.918693\n",
      "[36]\ttrain-auc:0.950979\teval-auc:0.918545\n",
      "[37]\ttrain-auc:0.951295\teval-auc:0.918206\n",
      "[38]\ttrain-auc:0.951497\teval-auc:0.918386\n",
      "[39]\ttrain-auc:0.952266\teval-auc:0.918386\n",
      "[40]\ttrain-auc:0.952787\teval-auc:0.918209\n",
      "[41]\ttrain-auc:0.954686\teval-auc:0.918919\n",
      "[42]\ttrain-auc:0.955148\teval-auc:0.918827\n",
      "[43]\ttrain-auc:0.955259\teval-auc:0.918913\n",
      "[44]\ttrain-auc:0.955888\teval-auc:0.918319\n",
      "[45]\ttrain-auc:0.956673\teval-auc:0.918894\n",
      "[46]\ttrain-auc:0.957416\teval-auc:0.918819\n",
      "[47]\ttrain-auc:0.957556\teval-auc:0.91893\n",
      "[48]\ttrain-auc:0.957683\teval-auc:0.918829\n",
      "[49]\ttrain-auc:0.958015\teval-auc:0.918937\n"
     ]
    }
   ],
   "source": [
    "# train model\n",
    "xgb_trained = xgb.train(params, dtrain, 50, watchlist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:35.263932Z",
     "start_time": "2018-12-07T18:03:35.259944Z"
    }
   },
   "outputs": [],
   "source": [
    "# define feature list\n",
    "features = [\"satisfaction_level\",\n",
    "            \"last_evaluation\",\n",
    "            \"number_project\",\n",
    "            \"average_montly_hours\",\n",
    "            \"time_spend_company\",\n",
    "            \"work_accident\",\n",
    "            \"promotion_last_5years\",\n",
    "            \"sales\",\n",
    "            \"salary\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:35.674789Z",
     "start_time": "2018-12-07T18:03:35.266001Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.03118302]\n",
      "Final logit contribution: -3.436201676579, predicted probability: 0.03118302986210538\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get a sample\n",
    "sample = dtest.slice([0])\n",
    "\n",
    "# model prediction\n",
    "print(xgb_trained.predict(sample))\n",
    "\n",
    "# visualize the feature contribution\n",
    "plt.figure(figsize=(10,10))\n",
    "plot_contribution(xgb_trained, sample, features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-12-07T18:03:35.931808Z",
     "start_time": "2018-12-07T18:03:35.676776Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.49110422]\n",
      "Final logit contribution: -0.03558703464999939, predicted probability: 0.4911041801506609\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get another sample\n",
    "sample = dtest.slice([1])\n",
    "\n",
    "# model prediction\n",
    "print(xgb_trained.predict(sample))\n",
    "\n",
    "# visualize the feature contribution\n",
    "plt.figure(figsize=(10,10))\n",
    "plot_contribution(xgb_trained, sample, features);"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}