DecisionTree_NaiveBayes.ipynb

{
 "metadata": {
  "name": "",
<<<<<<< HEAD
  "signature": "sha256:ada2302c0585f326cf46ad483c84ac80c0274fb322e4d9f1afdd83f658b5b3b1"
=======
  "signature": "sha256:47bd0cc6ae809cddfb0307e0a4e533c473d414131d0cc7018d208ded1d886910"
>>>>>>> d95a096f7680bf8c5e63879af592a9d868daa2d1
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def ROC(label,result):\n",
      "    %matplotlib inline\n",
      "    from sklearn.utils import shuffle\n",
      "    from sklearn.metrics import roc_curve, auc\n",
      "    import pylab as pl\n",
      "    import numpy as py\n",
      "    # Compute ROC curve and area the curve\n",
      "    Y = np.array(label)\n",
      "    fpr, tpr, thresholds = roc_curve(Y, result)\n",
      "    roc_auc = auc(fpr, tpr)\n",
      "    print \"Area under the ROC curve : %f\" % roc_auc\n",
      "\n",
      "    # Plot ROC curve\n",
      "    pl.clf()\n",
      "    pl.plot(fpr, tpr, label='ROC curve (area = %0.2f)' % roc_auc)\n",
      "    pl.plot([0, 1], [0, 1], 'k--')\n",
      "    pl.xlim([0.0, 1.0])\n",
      "    pl.ylim([0.0, 1.0])\n",
      "    pl.xlabel('False Positive Rate')\n",
      "    pl.ylabel('True Positive Rate')\n",
      "    pl.title('Receiver operating characteristic')\n",
      "    pl.legend(loc=\"lower right\")\n",
      "    pl.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np\n",
      "from sklearn import tree\n",
      "from sklearn import svm\n",
      "from sklearn.naive_bayes import GaussianNB\n",
      "\n",
      "# read in the data\n",
      "def readInData():\n",
      "    file = open('Data/train_1234_new_rehashed_balanced.txt')\n",
      "\n",
      "    trainData = []\n",
      "    trainLabel = []\n",
      "\n",
      "    # prepare the training data\n",
      "    for line in file:\n",
      "        li = list(map(int, line.split(',')))\n",
      "        trainLabel.append(li[0])\n",
      "        trainData.append(li[1:11])\n",
      "    \n",
      "    file = open('Data/test_6_new_rehashed.txt')\n",
      "    testData = []\n",
      "    testLabel = []\n",
      "    \n",
      "    for line in file:\n",
      "        li = list(map(int, line.split(',')))\n",
      "        testLabel.append(li[0])\n",
      "        testData.append(li[1:11])\n",
      "    \n",
      "    return (trainData, trainLabel, testData, testLabel)\n",
      "\n",
      "\n",
      "(trainData, trainLabel, testData, testLabel) = readInData()\n",
      "\n",
      "# train decision tree\n",
      "clf = tree.DecisionTreeClassifier(min_samples_split = 500, min_samples_leaf = 10)\n",
      "clf.fit(trainData, trainLabel)\n",
      "result = clf.predict_proba(testData)\n",
      "result = result[:, 1:2]\n",
      "\n",
      "ROC(testLabel, result)\n",
      "\n",
      "labels = clf.predict(testData)\n",
      "cnt = 0\n",
      "for i in range(0, len(testLabel)):\n",
      "    if labels[i] != testLabel[i]:\n",
      "        cnt = cnt + 1\n",
      "\n",
      "print float(cnt)/float(len(testLabel))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Area under the ROC curve : 0.694667\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEZCAYAAACTsIJzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npZRSSimllFJKKaWUUkoppcLF/wPQHliSp8Eq0QAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x13b679110>"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.477988913321\n"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.naive_bayes import GaussianNB\n",
      "# train naive bayes\n",
      "clfNB = GaussianNB()\n",
      "# clfNB = MultinomialNB()\n",
      "clfNB.fit(trainData, trainLabel)\n",
      "result = clfNB.predict_proba(testData)\n",
      "result = result[:, 1:2]\n",
      "\n",
      "ROC(testLabel, result)\n",
      "\n",
      "labels = clfNB.predict(testData)\n",
      "cnt = 0\n",
      "for i in range(0, len(testLabel)):\n",
      "    if labels[i] != testLabel[i]:\n",
      "        cnt = cnt + 1\n",
      "\n",
      "print float(cnt)/float(len(testLabel))\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Area under the ROC curve : 0.604054\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEZCAYAAACTsIJzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbY81nzTrAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x13b5bafd0>"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.612917556121\n"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.ensemble import RandomForestClassifier\n",
      "# train random forest\n",
      "clfRF = RandomForestClassifier(n_estimators = 20, min_samples_split = 500, min_samples_leaf = 10, n_jobs = 2)\n",
      "clfRF.fit(trainData, trainLabel)\n",
      "result = clfRF.predict_proba(testData)\n",
      "result = result[:, 1:2]\n",
      "\n",
      "ROC(testLabel, result)\n",
      "\n",
      "labels = clfRF.predict(testData)\n",
      "cnt = 0\n",
      "for i in range(0, len(testLabel)):\n",
      "    if labels[i] != testLabel[i]:\n",
      "        cnt = cnt + 1\n",
      "\n",
      "print float(cnt)/float(len(testLabel))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Area under the ROC curve : 0.695860\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEZCAYAAACTsIJzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x13b6aa410>"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.454658301285\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np\n",
      "import math\n",
      "def featureExtract(data, label):\n",
      "    clickCount = {}\n",
      "    count = {}\n",
      "    for i in range(0, len(data)):\n",
      "        if data[i] not in count:\n",
      "            count[data[i]] = 1\n",
      "            clickCount[data[i]] = 0\n",
      "        else:\n",
      "            count[data[i]] = count[data[i]] + 1\n",
      "        \n",
      "        clickCount[data[i]] = clickCount[data[i]] + label[i]\n",
      "    \n",
      "    return (count, clickCount)\n",
      "\n",
      "def featureTransfrom(data, start, label):\n",
      "    nFeature = np.shape(data)[1]\n",
      "    tData = []\n",
      "    tData.append(data[:, 0])\n",
      "#     tData = data[:, 0]\n",
      "#     tData = np.concatenate((tData, data[:,1]), axis = 1)\n",
      "    for i in range(start, nFeature):\n",
      "        subData = data[:, i]\n",
      "        subData = subData.tolist()\n",
      "        (count, clickCount) = featureExtract(subData, label)\n",
      "        rate = {}\n",
      "        for (k, v) in count.items():\n",
      "            if clickCount[k] == 0:\n",
      "                rate[k] = 0\n",
      "            else:\n",
      "                rate[k] = float(clickCount[k]) / float(v)\n",
      "        \n",
      "        li = []\n",
      "        for j in range(0, np.shape(subData)[0]):\n",
      "            li.append(rate[subData[j]])\n",
      "        li = np.array(li)\n",
      "        li = np.transpose(li)\n",
      "        tData.append(li)\n",
      "    \n",
      "    tData = np.array(tData)\n",
      "    tData = np.transpose(tData)\n",
      "    return tData\n",
      "\n",
      "def featureExtraTest(data, start):\n",
      "    tData = []\n",
      "    tData.append(data[:, 0])\n",
      "    nFeature = np.shape(data)[1]\n",
      "    nData = np.shape(data)[0]\n",
      "    for i in range(start, nFeature):\n",
      "        count = {}\n",
      "        subData = data[:, i]\n",
      "        subData = subData.tolist()\n",
      "        for j in range(0, nData):\n",
      "            if subData[j] not in count:\n",
      "                count[subData[j]] = 1\n",
      "            else:\n",
      "                count[subData[j]] = count[subData[j]] + 1\n",
      "        rate = {}\n",
      "        for (k, v) in count.items():\n",
      "            rate[k] = float(v) / float(nData)\n",
      "            \n",
      "        li = []\n",
      "        for j in range(0, nData):\n",
      "            li.append(rate[subData[j]])\n",
      "        tData.append(li)\n",
      "    \n",
      "    tData = np.array(tData)\n",
      "    tData = np.transpose(tData)\n",
      "    return tData"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 36
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np\n",
      "from sklearn.linear_model import LinearRegression as Lr\n",
      "from sklearn.naive_bayes import GaussianNB\n",
      "\n",
      "(trainData, trainLabel, testData, testLabel) = readInData()\n",
      "    \n",
      "tData = featureTransfrom(np.array(trainData), 1, np.array(trainLabel))\n",
      "\n",
      "gnb = GaussianNB()\n",
      "gnb.fit(tData, trainLabel)\n",
      "\n",
      "# lrInstance = Lr()\n",
      "# lrInstance.fit(tData, trainLabel)\n",
      "\n",
      "tData = featureExtraTest(np.array(testData), 1)\n",
      "for T in range(0, 20):\n",
      "    tResult = lrInstance.predict(tData)\n",
      "    tData = featureTransfrom(np.array(testData), 1, tResult)\n",
      "\n",
      "ROC(testLabel, tResult)\n",
      "\n",
      "tData = featureTransfrom(np.array(trainData), 1, np.array(trainLabel))\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Area under the ROC curve : 0.632438\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEZCAYAAACTsIJzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nH0BIA7Lt2E5qAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10590d810>"
       ]
      }
     ],
     "prompt_number": 41
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sys\n",
      "sys.path.append('/Users/peiyang/Documents/Class/CS260/libsvm-3.20/python')\n",
      "\n",
      "from svmutil import *\n",
      "\n",
      "labels1 = clf.predict_proba(trainData)\n",
      "labels1 = labels1[:, 1:2]\n",
      "labels2 = clfRF.predict_proba(trainData)\n",
      "labels2 = labels2[:, 1:2]\n",
      "labels3 = clfNB.predict_proba(trainData)\n",
      "labels3 = labels3[:, 1:2]\n",
      "\n",
      "newInputs = np.concatenate((labels1, labels2, labels3), axis = 1)\n",
      "newList = newInputs.tolist()\n",
      "model = svm_train(trainLabel, newList, '-c 1')\n",
      "\n",
      "p_labels, p_acc, p_vals = svm_predict(testLabel, testData, model)\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "AssertionError",
       "evalue": "",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-10-d2d5395f3398>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     12\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0mnewInputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconcatenate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlabels1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msvm_train\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainLabel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnewInputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'-c 1'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m \u001b[0mp_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp_acc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp_vals\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msvm_predict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtestLabel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtestData\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/peiyang/Documents/Class/CS260/libsvm-3.20/python/svmutil.pyc\u001b[0m in \u001b[0;36msvm_train\u001b[0;34m(arg1, arg2, arg3)\u001b[0m\n\u001b[1;32m    123\u001b[0m         \u001b[0mprob\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparam\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 125\u001b[0;31m                 \u001b[0;32massert\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    126\u001b[0m                 \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marg1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg3\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    127\u001b[0m                 \u001b[0mparam\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msvm_parameter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moptions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mAssertionError\u001b[0m: "
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "model = svm_train(trainLabel, newList, '-c 1')\n",
      "\n",
      "p_labels, p_acc, p_vals = svm_predict(testLabel, testData, model)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print len(newList)\n",
      "print len(trainLabel)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "5609172\n"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "p_labels"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}