xgboost-regression-boston-housing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import xgboost as xb\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_boston"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = load_boston()\n",
    "X = data.data\n",
    "y = data.target\n",
    "cols = data.feature_names"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Split data in test & train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "trainX, testX, trainY, testY = train_test_split(X,y)\n",
    "\n",
    "dTrainMatrix = xb.DMatrix(trainX, trainY)\n",
    "dTestMatrix = xb.DMatrix(testX, testY)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train the Model and Predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {\"learning_rate\":\"0.9\", \"objective\":\"reg:linear\",\"n_estimators\":\"10\"}\n",
    "xb_reg_cl = xb.train(params=params, dtrain=dTrainMatrix)\n",
    "predictedY = xb_reg_cl.predict(data=dTestMatrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Calculate RMSE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.951757692001573\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_squared_error\n",
    "rmse = np.sqrt(mean_squared_error(testY, predictedY))\n",
    "print (rmse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tune Regularisation parameters \n",
    "## Trying L1 regularisation\n",
    "\n",
    "\t- xgboost.cv() would train the model but it does not return the model. It is typically used to find the righ parameters and once those are known we can do regular xgboost.train() and xgboost.predict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.7780018196391573, 1.0095510916079484, 0.9413793071226924, 1.6799027036935918]\n"
     ]
    }
   ],
   "source": [
    "dTrainMatrix = xb.DMatrix(trainX, trainY)\n",
    "dTestMatrix = xb.DMatrix(testX, testY)\n",
    "\n",
    "params = {\"learning_rate\":\"0.9\", \"objective\":\"reg:linear\",\"n_estimators\":\"10\"}\n",
    "\n",
    "l1_params = [1,10,100,1000]\n",
    "rmse = []\n",
    "for l1 in l1_params:\n",
    "    params['alpha']=l1\n",
    "    cv_results = xb.cv(params=params, dtrain=dTrainMatrix, num_boost_round=5, nfold=5, metrics='rmse', as_pandas=True)\n",
    "    rmse.append(cv_results['test-rmse-std'][4])\n",
    "\n",
    "print (rmse)\n",
    "\n",
    "# We see that the RMSE is least when Alpha is 1, so we can use that value\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.515664951893516\n"
     ]
    }
   ],
   "source": [
    "params = {\"learning_rate\":\"0.9\", \"objective\":\"reg:linear\",\"n_estimators\":\"10\",\"alpha\":\"1\"}\n",
    "model = xb.train(params=params, dtrain=dTrainMatrix, num_boost_round=5)\n",
    "predictedY = model.predict(data=dTestMatrix)\n",
    "\n",
    "rmse = np.sqrt(mean_squared_error(testY, predictedY))\n",
    "print (rmse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot Feature Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x17eb0c05eb8>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "xb.plot_importance(model)"
   ]
  },
  {
   "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
}