Forest_Regressor.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import figure\n",
    "import matplotlib.lines as mlines\n",
    "import matplotlib.transforms as mtransforms\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import sklearn\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error, explained_variance_score\n",
    "def mean_absolute_percentage_error(y_true, y_pred): \n",
    "    y_true, y_pred = np.array(y_true), np.array(y_pred)\n",
    "    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"data/data_cleaned.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# one hot\n",
    "data = pd.get_dummies(data, prefix = [\"HEAT\", \"AC\", \"STYLE\", \"STRUCT\", \"GRADE\", \"CNDTN\", \"EXTWALL\", \"INTWALL\", \"ROOF\", \"ASSESSMENT_SUBNBHD\"], columns = [\"HEAT\", \"AC\", \"STYLE\", \"STRUCT\", \"GRADE\", \"CNDTN\", \"EXTWALL\", \"INTWALL\", \"ROOF\", \"ASSESSMENT_SUBNBHD\"])\n",
    "# split\n",
    "x_data = data.drop([\"PRICE\"], axis=1)\n",
    "y_data = data[\"PRICE\"]\n",
    "# normalization\n",
    "x_data = x_data.astype(float)\n",
    "x_data = x_data.apply(lambda x: (x - x.min()) / (x.max() - x.min()))\n",
    "# train & test\n",
    "x_train, x_test, y_train, y_test = train_test_split(x_data, y_data, test_size=0.33, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest Regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.86607619253193824"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# create an fit random forest\n",
    "model = RandomForestRegressor(n_estimators=500, n_jobs=-1, oob_score=True)\n",
    "model.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2 Score: 0.866076192532\n",
      "Explained variance score: 0.866077802603\n",
      "Mean absolute error: 90309.2684079\n",
      "Mean absolute percentage error: 22.8124511462\n",
      "Mean squared error: 27698156792.9\n"
     ]
    }
   ],
   "source": [
    "print(\"R2 Score: \" + str(r2_score(y_test, y_pred)))\n",
    "print(\"Explained variance score: \" + str(explained_variance_score(y_test, y_pred)))\n",
    "print(\"Mean absolute error: \" + str(mean_absolute_error(y_test, y_pred)))\n",
    "print(\"Mean absolute percentage error: \" + str(mean_absolute_percentage_error(y_test, y_pred)))\n",
    "print(\"Mean squared error: \" + str(mean_squared_error(y_test, y_pred)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEDCAYAAAArwUMAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBEEQBEf+P4Z/2gWvqNqvAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1f401008e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.scatter(y_test, y_pred, marker='x')\n",
    "plt.xlim(0, 5e6)\n",
    "plt.ylim(0, 1e7)\n",
    "ax.plot([0, y_test.max()], [0, y_test.max()], 'r--', lw=4)\n",
    "plt.show()  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred_train = model.predict(x_train)\n",
    "pd.DataFrame(pred_train).to_csv(\"predictions/forest.csv\", index=False)\n",
    "pd.DataFrame(y_pred).to_csv(\"predictions/forest_test.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pd.DataFrame(y_train).to_csv(\"predictions/y.csv\", index=False)\n",
    "pd.DataFrame(y_test).to_csv(\"predictions/y_test.csv\", index=False)"
   ]
  },
  {
   "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}