Linear Regression.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "\n",
    "from numpy import genfromtxt\n",
    "from sklearn.datasets import load_boston"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def read_dataset(filePath,delimiter=','):\n",
    "    return genfromtxt(filePath, delimiter=delimiter)\n",
    "\n",
    "def read_boston_data():\n",
    "    boston = load_boston()\n",
    "    features = np.array(boston.data)\n",
    "    labels = np.array(boston.target)\n",
    "    return features, labels\n",
    "\n",
    "def feature_normalize(dataset):\n",
    "    mu = np.mean(dataset,axis=0)\n",
    "    sigma = np.std(dataset,axis=0)\n",
    "    return (dataset - mu)/sigma\n",
    "\n",
    "def append_bias_reshape(features,labels):\n",
    "    n_training_samples = features.shape[0]\n",
    "    n_dim = features.shape[1]\n",
    "    f = np.reshape(np.c_[np.ones(n_training_samples),features],[n_training_samples,n_dim + 1])\n",
    "    l = np.reshape(labels,[n_training_samples,1])\n",
    "    return f, l"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "features,labels = read_boston_data()\n",
    "normalized_features = feature_normalize(features)\n",
    "f, l = append_bias_reshape(normalized_features,labels)\n",
    "n_dim = f.shape[1]\n",
    "\n",
    "rnd_indices = np.random.rand(len(f)) < 0.80\n",
    "\n",
    "train_x = f[rnd_indices]\n",
    "train_y = l[rnd_indices]\n",
    "test_x = f[~rnd_indices]\n",
    "test_y = l[~rnd_indices]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "learning_rate = 0.01\n",
    "training_epochs = 1000\n",
    "cost_history = np.empty(shape=[1],dtype=float)\n",
    "\n",
    "X = tf.placeholder(tf.float32,[None,n_dim])\n",
    "Y = tf.placeholder(tf.float32,[None,1])\n",
    "W = tf.Variable(tf.ones([n_dim,1]))\n",
    "\n",
    "init = tf.initialize_all_variables()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_ = tf.matmul(X, W)\n",
    "cost = tf.reduce_mean(tf.square(y_ - Y))\n",
    "training_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 22.2268\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAD7CAYAAACMlyg3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKzEnATOzEnMSMDMrMScBM7MS+3/MKbAwopZ4LwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd18d872b50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAEPCAYAAACgFqixAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd18d63b5d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "with tf.Session() as sess:\n",
    "    sess.run(init)\n",
    "    for epoch in range(training_epochs):\n",
    "        sess.run(training_step,feed_dict={X:train_x,Y:train_y})\n",
    "        cost_history = np.append(cost_history,sess.run(cost,feed_dict={X: train_x,Y: train_y}))\n",
    "          \n",
    "    #calculate mean square error \n",
    "    pred_y = sess.run(y_, feed_dict={X: test_x})\n",
    "    mse = tf.reduce_mean(tf.square(pred_y - test_y))\n",
    "    print(\"MSE: %.4f\" % sess.run(mse)) \n",
    "    \n",
    "    #plot cost\n",
    "    plt.plot(range(len(cost_history)),cost_history)\n",
    "    plt.axis([0,training_epochs,0,np.max(cost_history)])\n",
    "    plt.show()\n",
    "    \n",
    "    fig, ax = plt.subplots()\n",
    "    ax.scatter(test_y, pred_y)\n",
    "    ax.plot([test_y.min(), test_y.max()], [test_y.min(), test_y.max()], 'k--', lw=3)\n",
    "    ax.set_xlabel('Measured')\n",
    "    ax.set_ylabel('Predicted')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}