annotate notebooks/explain_components.ipynb @ 9:c4841876a8ff branch-tests

adding notebooks and trying to explain classifier coefficients

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LDA\n",
    "\n",
    "import sys\n",
    "sys.path.append('../')\n",
    "import scripts.map_and_average as mapper\n",
    "import scripts.util_feature_learning as util_feature_learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/import/c4dm-04/mariap/train_data_melodia_8.pickle\n"
     ]
    },
    {
     "ename": "IOError",
     "evalue": "[Errno 2] No such file or directory: '/import/c4dm-04/mariap/train_data_melodia_8.pickle'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIOError\u001b[0m                                   Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-8-aa3c9e978b25>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mtrainset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtestset\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmapper\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_train_val_test_sets\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/Users/mariapanteli/Documents/QMUL/Code/MyPythonCode/plosone_underreview/scripts/map_and_average.pyc\u001b[0m in \u001b[0;36mload_train_val_test_sets\u001b[0;34m()\u001b[0m\n\u001b[1;32m     69\u001b[0m     '''\n\u001b[1;32m     70\u001b[0m     \u001b[0;32mprint\u001b[0m \u001b[0mINPUT_FILES\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 71\u001b[0;31m     \u001b[0mtrainset\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_data_from_pickle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mINPUT_FILES\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\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     72\u001b[0m     \u001b[0mvalset\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_data_from_pickle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mINPUT_FILES\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     73\u001b[0m     \u001b[0mtestset\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_data_from_pickle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mINPUT_FILES\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/Users/mariapanteli/Documents/QMUL/Code/MyPythonCode/plosone_underreview/scripts/map_and_average.pyc\u001b[0m in \u001b[0;36mload_data_from_pickle\u001b[0;34m(pickle_file)\u001b[0m\n\u001b[1;32m     56\u001b[0m     '''load frame based features and labels from pickle file\n\u001b[1;32m     57\u001b[0m     '''\n\u001b[0;32m---> 58\u001b[0;31m     \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpickle_file\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'rb'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     59\u001b[0m         \u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maudiolabels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpickle\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     60\u001b[0m     \u001b[0;31m# remove 'unknown' and 'unidentified' country\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mIOError\u001b[0m: [Errno 2] No such file or directory: '/import/c4dm-04/mariap/train_data_melodia_8.pickle'"
     ]
    }
   ],
   "source": [
    "trainset, valset, testset = mapper.load_train_val_test_sets()\n",
    "traindata, trainlabels, trainaudiolabels = trainset\n",
    "valdata, vallabels, valaudiolabels = valset\n",
    "testdata, testlabels, testaudiolabels = testset\n",
    "labels = np.concatenate((trainlabels, vallabels, testlabels)).ravel()\n",
    "audiolabels = np.concatenate((trainaudiolabels, valaudiolabels, testaudiolabels)).ravel()\n",
    "print traindata.shape, valdata.shape, testdata.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## explain LDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "min_variance = 0.99\n",
    "feat_labels, feat_inds = mapper.get_feat_inds(n_dim=traindata.shape[1])\n",
    "for i in range(len(feat_inds)):\n",
    "    print \"mapping \" + feat_labels[i]\n",
    "    inds = feat_inds[i]\n",
    "    ssm_feat = util_feature_learning.Transformer()\n",
    "    if min_variance is not None:\n",
    "        ssm_feat.fit_data(traindata[:, inds], trainlabels, n_components=len(inds), pca_only=True)\n",
    "        n_components = np.where(ssm_feat.pca_transformer.explained_variance_ratio_.cumsum()>min_variance)[0][0]+1\n",
    "        print n_components, len(inds)\n",
    "        ssm_feat.fit_lda_data(traindata[:, inds], trainlabels, n_components=n_components)\n",
    "\n",
    "        WW = ssm_feat.lda_transformer.scalings_\n",
    "        plt.figure()\n",
    "        plt.imshow(WW[:, :n_components].T, aspect='auto')\n",
    "        plt.colorbar()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## explain classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_list, Y, Yaudio = pickle.load(open('../data/lda_data_melodia_8.pickle','rb'))\n",
    "Xrhy, Xmel, Xmfc, Xchr = X_list\n",
    "X = np.concatenate((Xrhy, Xmel, Xmfc, Xchr), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ssm_feat.classify_and_save(X_train, Y_train, X_test, Y_test, transform_label=\" \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def components_plot(lda_transformer, XX, n_comp=42, figurename=None):\n",
    "    WW=lda_transformer.scalings_\n",
    "    Xlda=lda_transformer.transform(XX)\n",
    "    Xww=numpy.dot(XX, WW[:, :n_comp])\n",
    "    plt.figure()\n",
    "    plt.imshow(Xlda - Xww, aspect='auto')\n",
    "    plt.figure()\n",
    "    plt.imshow(Xlda, aspect='auto')\n",
    "    plt.figure()\n",
    "    plt.imshow(Xww, aspect='auto')\n",
    "    plt.figure()\n",
    "    plt.imshow(WW[:, :n_comp], aspect='auto') # this explains the weights up to n_components=64\n",
    "    if figurename is not None:\n",
    "        plt.savefig(figurename)\n",
    "\n",
    "XX = traindata[:, inds]\n",
    "components_plot(ssm_feat.lda_transformer, XX, n_comp=n_components)"
   ]
  }
 ],
 "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": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}