linear_regression.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoca 0: 2069.6319333978354\n",
      "Epoca 1: 2117.0123581953535\n",
      "Epoca 2: 2092.302723001866\n",
      "Epoca 3: 2068.5080461938464\n",
      "Epoca 4: 2045.591184088162\n",
      "Epoca 5: 2023.5146448101316\n",
      "Epoca 6: 2002.2447619835536\n",
      "Epoca 7: 1981.748338803649\n",
      "Epoca 8: 1961.9944411260742\n",
      "Epoca 9: 1942.9520116143283\n",
      "Epoca 10: 1924.5930823644712\n",
      "Epoca 11: 1906.8898800636332\n",
      "Epoca 12: 1889.8164505837929\n",
      "Epoca 13: 1873.347133841543\n",
      "Epoca 14: 1857.4588400604468\n",
      "Epoca 15: 1842.1278742424079\n",
      "Epoca 16: 1827.332495119955\n",
      "Epoca 17: 1813.0520579712022\n",
      "Epoca 18: 1799.2660847636982\n",
      "Epoca 19: 1785.9562132299961\n",
      "Epoca 20: 1773.1024853109072\n",
      "Epoca 21: 1760.689129482884\n",
      "Epoca 22: 1748.6984157081515\n",
      "Epoca 23: 1737.1138680398553\n",
      "Epoca 24: 1725.920873066732\n",
      "Epoca 25: 1715.1046249579008\n",
      "Epoca 26: 1704.6500954309377\n",
      "Epoca 27: 1694.5447134910141\n",
      "Epoca 28: 1684.7746311347667\n",
      "Epoca 29: 1675.328450968245\n",
      "Epoca 30: 1666.1935385839038\n",
      "Epoca 31: 1657.3584002084322\n",
      "Epoca 32: 1648.8122658529207\n",
      "Epoca 33: 1640.5440742547091\n",
      "Epoca 34: 1632.5446836102221\n",
      "Epoca 35: 1624.8043315147183\n",
      "Epoca 36: 1617.3126799958602\n",
      "Epoca 37: 1610.0622532456405\n",
      "Epoca 38: 1603.0433557207386\n",
      "Epoca 39: 1596.2479176106197\n",
      "Epoca 40: 1589.668056331575\n",
      "Epoca 41: 1583.2965242617897\n",
      "Epoca 42: 1577.126371285745\n",
      "Epoca 43: 1571.1501190634\n",
      "Epoca 44: 1565.360979151513\n",
      "Epoca 45: 1559.7523780798629\n",
      "Epoca 46: 1554.3184364555138\n",
      "Epoca 47: 1549.0529469620615\n",
      "Epoca 48: 1543.950059985476\n",
      "Epoca 49: 1539.0050282141283\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb7e2ae3fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import xlrd\n",
    "\n",
    "import utils\n",
    "\n",
    "DATA_FILE = \"data/fire_theft.xls\"\n",
    "\n",
    "#read data\n",
    "book = xlrd.open_workbook(DATA_FILE, encoding_override=\"utf-8\")\n",
    "sheet = book.sheet_by_index(0)\n",
    "data = np.asarray([sheet.row_values(i) for i in range(1, sheet.nrows)])\n",
    "n_samples = sheet.nrows-1\n",
    "\n",
    "#create placeholders (X) (Y)\n",
    "X = tf.placeholder(tf.float32, name=\"X\")\n",
    "Y = tf.placeholder(tf.float32, name=\"Y\")\n",
    "\n",
    "#create weight, bias, initialized to 0\n",
    "w = tf.Variable(0.0, name=\"weight\", dtype=tf.float32)\n",
    "b = tf.Variable(0.0, name=\"bias\", dtype=tf.float32)\n",
    "\n",
    "#predict Y from X\n",
    "Y_predicted = X*w + b\n",
    "\n",
    "#use square root error as a loss function\n",
    "loss = tf.square(Y - Y_predicted, name=\"loss\")\n",
    "#loss = utils.hubber_loss(Y - Y_predicted)\n",
    "\n",
    "#use gradient descent + learn_rate 0.01 to minimize loss\n",
    "optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001).minimize(loss)\n",
    "\n",
    "#train model\n",
    "with tf.Session() as sess:\n",
    "    #initialize all variables\n",
    "    sess.run(tf.global_variables_initializer()) \n",
    "    writer = tf.summary.FileWriter(\"./graphs/linearReg\", sess.graph)\n",
    "    #train model\n",
    "    for i in range(50):\n",
    "        total_loss = 0\n",
    "        for x,y in data:\n",
    "            #session run optimizer to minimize loss + fetch data\n",
    "            _, l = sess.run([optimizer, loss], feed_dict={X: x, Y: y})\n",
    "            total_loss += l\n",
    "        print(\"Epoca {0}: {1}\".format(i, total_loss/n_samples))\n",
    "    writer.close()\n",
    "    \n",
    "    w, b = sess.run([w, b])\n",
    "    \n",
    "X, Y = data.T[0], data.T[1]\n",
    "plt.plot(X, Y, \"bo\", label=\"Real data\")\n",
    "plt.plot(X, X*w+b, \"r\", label=\"Predicted data\")\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}