16B-Classification-III-Random-Projection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Projections Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQoAAAEQCAYAAABIhjo0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGRSSsv4fLKlarTNr0OEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x106388208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import datasets\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#Load the digits dataset\n",
    "digits = datasets.load_digits()\n",
    "\n",
    "#Display the first digit\n",
    "plt.figure(4, figsize=(3, 3))\n",
    "plt.imshow(digits.images[2], cmap=plt.cm.gray_r, interpolation='nearest')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/crovella/anaconda3/lib/python3.5/site-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.random_projection import SparseRandomProjection\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.linear_model import Perceptron\n",
    "from sklearn.linear_model import PassiveAggressiveClassifier\n",
    "from sklearn.naive_bayes import BernoulliNB, MultinomialNB\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.neighbors import NearestCentroid\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn import tree\n",
    "from sklearn import metrics\n",
    "from sklearn import datasets\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "accuracies = []\n",
    "components = np.int32(np.linspace(2, 64, 20))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "digits = datasets.load_digits()\n",
    "split = train_test_split(digits.data, digits.target, test_size = 0.3,\n",
    "    random_state = 42)\n",
    "(trainData, testData, trainTarget, testTarget) = split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.99259259259259258"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#model = tree.DecisionTreeClassifier()\n",
    "#model.fit(trainData, trainTarget)\n",
    "\n",
    "model = KNeighborsClassifier()\n",
    "model.fit(trainData, trainTarget)\n",
    "\n",
    "#model = LinearSVC()\n",
    "#model.fit(trainData, trainTarget)\n",
    "baseline = metrics.accuracy_score(model.predict(testData), testTarget)\n",
    "baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "# loop over the projection sizes\n",
    "for comp in components:\n",
    "    # create the random projection\n",
    "    sp = SparseRandomProjection(n_components = comp)\n",
    "    X = sp.fit_transform(trainData)\n",
    "    #print X.shape\n",
    " \n",
    "    # train a classifier on the sparse random projection\n",
    "    #model = tree.DecisionTreeClassifier()\n",
    "    #model = LinearSVC()\n",
    "    model = KNeighborsClassifier()\n",
    "    model.fit(X, trainTarget)\n",
    " \n",
    "    # evaluate the model and update the list of accuracies\n",
    "    test = sp.transform(testData)\n",
    "    accuracies.append(metrics.accuracy_score(model.predict(test), testTarget))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhUAAAGSCAYAAABUuTODAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npFIYKiRJUin+PxOgvSEVwzyOAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x106370fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# create the figure\n",
    "plt.figure()\n",
    "plt.suptitle(\"Accuracy of Sparse Projection on Digits\")\n",
    "plt.xlabel(\"# of Components\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.xlim([2, 64])\n",
    "plt.ylim([0, 1.0])\n",
    " \n",
    "# plot the baseline and random projection accuracies\n",
    "plt.plot(components, [baseline] * len(accuracies), color = \"r\")\n",
    "plt.plot(components, accuracies)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "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": 0
}