neural_nets_with_radio.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Deep Learning with Radio Ineterferometers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook, we are trying to use radio visibilities and neural networks to predict if an antenna will be good or bad. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import glob\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Get all the features in a condensed form.\n",
    "# d1 = np.loadtxt('data/stats_1e2.txt')\n",
    "# ants = d1[:,0]#antenna numbers\n",
    "# f1 = d1[:,-1]#standard devs of solved delays\n",
    "# m,v,f = capo.miriad.read_files(['zen.2456679.43314.xx.uvcRRE'],'auto','xx',verbose=True)\n",
    "# bad_antennae = [8,14,15,16,26,27,28,34,37,38,42,46,50,53,72,74,82,84,85,110]\n",
    "\n",
    "# feature_dict = {}\n",
    "# times = m['times']\n",
    "# for a,ff in zip(ants,f1):\n",
    "#    ant_flag = 1\n",
    "#    if a in bad_antennae: \n",
    "#        print 'flagging: ',a \n",
    "#        ant_flag = 0\n",
    "#    for samp,tsample in enumerate(times):\n",
    "#     feature_dict[a] = [np.hstack((ff,np.real(v[(a,a)]['xx'][time]),ant_flag))]\n",
    "#        feature_dict[str(a) + str(tsample)] =list(np.hstack((np.real(v[(a,a)]['xx'][samp]),ant_flag)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pos_sample_dir = 'data/good'\n",
    "neg_sample_dir = 'data/bad'\n",
    "pos_sample_files = [np.load(f) for f in glob.glob(pos_sample_dir+'/*data.npz')]\n",
    "neg_sample_files = [np.load(f) for f in glob.glob(neg_sample_dir+'/*data.npz')]\n",
    "\n",
    "pos_sample_files2 = [np.load(f) for f in glob.glob(pos_sample_dir+'2/*.npz')]\n",
    "neg_sample_files2 = [np.load(f) for f in glob.glob(neg_sample_dir+'2/*.npz')]\n",
    "\n",
    "\n",
    "\n",
    "pos_samples = np.concatenate([np.real(d['arr_0']) for d in pos_sample_files])\n",
    "neg_samples = np.concatenate([np.real(d['arr_0']) for d in neg_sample_files])\n",
    "pos_samples2 = np.concatenate([np.real(d['arr_0']) for d in pos_sample_files2])\n",
    "neg_samples2 = np.concatenate([np.real(d['arr_0']) for d in neg_sample_files2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#need to normalize features\n",
    "Y = np.concatenate([np.ones(pos_samples.shape[0]),np.zeros(neg_samples.shape[0])])\n",
    "X = np.concatenate([pos_samples,neg_samples])\n",
    "xshape = X.shape\n",
    "X = X.reshape(xshape[0],1,xshape[1],xshape[2])\n",
    "X.shape\n",
    "#shuffle\n",
    "samps = np.arange(X.shape[0])\n",
    "np.random.shuffle(samps)\n",
    "X = X[samps]\n",
    "Y = Y[samps]\n",
    "ntrains = int(len(samps)/1.25)\n",
    "Xtrain = X[:ntrains]; ytrain = Y[:ntrains]\n",
    "Xtest = X[ntrains:]; ytest = Y[ntrains:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-1, 10)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABI4AAAE4CAYAAAA9wLhWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAgAAAKBLcAQAAABA1/8D0iiMOwXpEWEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2ac3c1ad5c90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,5))\n",
    "plt.plot(ytrain)\n",
    "plt.plot(ytest + 5)\n",
    "plt.ylim(-1,10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x2ac44be97a90>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHsAAAImCAYAAADDkYb9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZQ8AAADAQv4CYBq9NWcnXFwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2ac44bde2ed0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,10))\n",
    "plt.subplot(121)\n",
    "plt.imshow(Xtrain[np.where(ytrain==0)[0][0]].reshape(40,40))\n",
    "plt.title('Good Antenna')\n",
    "plt.subplot(122)\n",
    "plt.imshow(Xtrain[np.where(ytrain==1)[0][0]].reshape(40,40))\n",
    "plt.title('Bad Antenna')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to normalize all the images to have the same weights. We do this by subtracting the mean and dividing by the standard deviation, giving us a mean zero and variance 1 feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10.9802 5.40234\n",
      "-1.90139e-07 1.0\n"
     ]
    }
   ],
   "source": [
    "throwaway = Xtrain[5]\n",
    "throwaway.shape\n",
    "mn = np.mean(throwaway)\n",
    "vr = np.std(throwaway)\n",
    "throwaway = (throwaway -mn)/vr\n",
    "print mn,vr\n",
    "print np.mean(throwaway), np.var(throwaway)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "trainmeans = []\n",
    "trainstdevs = []\n",
    "testmeans = []\n",
    "teststdevs = []\n",
    "Xtrain_norm = np.zeros_like(Xtrain)\n",
    "Xtest_norm = np.zeros_like(Xtest)\n",
    "for xx in range(Xtrain.shape[0]):\n",
    "    trainmeans.append(np.mean(Xtrain[xx]))\n",
    "    trainstdevs.append(np.std(Xtrain[xx]))\n",
    "    Xtrain_norm[xx] = (Xtrain[xx] - trainmeans[-1])/trainstdevs[-1]\n",
    "for xx in range(Xtest.shape[0]):\n",
    "    testmeans.append(np.mean(Xtest[xx]))\n",
    "    teststdevs.append(np.std(Xtest[xx]))\n",
    "    Xtest_norm[xx] = (Xtest[xx] - testmeans[-1])/teststdevs[-1]\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Y2 = np.concatenate([np.ones(pos_samples2.shape[0]),np.ones(neg_samples2.shape[0])])\n",
    "#X2 = np.concatenate([pos_samples2,neg_samples2])\n",
    "#x2shape = X2.shape\n",
    "#X2 = X2.reshape(x2shape[0],1,x2shape[1],x2shape[2])\n",
    "##shuffle\n",
    "#samps = np.arange(X2.shape[0])\n",
    "#np.random.shuffle(samps)\n",
    "#Xtest = X2[samps][:30000]\n",
    "#ytest = Y2[samps][:30000]\n",
    "#Xtrain = X[:len(samps)/3]; ytrain = Y[:len(samps)/3]\n",
    "#Xtest = X[len(samps)/3:]; ytest = Y[len(samps)/3:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(121344, 1, 40, 40) (121344,)\n",
      "(30336, 1, 40, 40) (30336,)\n"
     ]
    }
   ],
   "source": [
    "print Xtrain_norm.shape, ytrain.shape\n",
    "print Xtest_norm.shape,  ytest.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Xtrain shape:', (121344, 1, 40, 40))\n",
      "(121344, 'train samples')\n",
      "(30336, 'test samples')\n",
      "Not using data augmentation.\n",
      "Train on 121344 samples, validate on 30336 samples\n",
      "Epoch 1/1\n",
      "121344/121344 [==============================] - 2729s - loss: 0.5441 - acc: 0.7768 - val_loss: 0.5508 - val_acc: 0.7835\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using Theano backend.\n"
     ]
    }
   ],
   "source": [
    "'''Train a simple deep CNN on the CIFAR10 small images dataset.\n",
    "GPU run command:\n",
    "    THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python cifar10_cnn.py\n",
    "It gets down to 0.65 test logloss in 25 epochs, and down to 0.55 after 50 epochs.\n",
    "(it's still underfitting at that point, though).\n",
    "Note: the data was pickled with Python 2, and some encoding issues might prevent you\n",
    "from loading it in Python 3. You might have to load it in Python 2,\n",
    "save it in a different format, load it in Python 3 and repickle it.\n",
    "'''\n",
    "\n",
    "#from __future__ import print_function\n",
    "from keras.datasets import cifar10\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras.layers import Convolution2D, MaxPooling2D\n",
    "from keras.optimizers import SGD\n",
    "from keras.utils import np_utils\n",
    "\n",
    "batch_size = 2048\n",
    "nb_classes = 2\n",
    "nb_epoch = 1\n",
    "data_augmentation = False#True\n",
    "\n",
    "# input image dimensions\n",
    "img_rows, img_cols = 40, 40\n",
    "# the CIFAR10 images are RGB\n",
    "img_channels = 1\n",
    "\n",
    "# the data, shuffled and split between train and test sets\n",
    "print('Xtrain shape:', Xtrain_norm.shape)\n",
    "print(Xtrain_norm.shape[0], 'train samples')\n",
    "print(Xtest_norm.shape[0], 'test samples')\n",
    "\n",
    "# convert class vectors to binary class matrices\n",
    "Ytrain = np_utils.to_categorical(ytrain, nb_classes)\n",
    "Ytest = np_utils.to_categorical(ytest, nb_classes)\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Convolution2D(32, 3, 3, border_mode='same',\n",
    "                        input_shape=(img_channels, img_rows, img_cols)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Convolution2D(32, 3, 3))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "#model.add(Droptout(0.25))\n",
    "model.add(Dropout(0.5))\n",
    "\n",
    "model.add(Convolution2D(64, 3, 3, border_mode='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Convolution2D(64, 3, 3))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.7))\n",
    "\n",
    "model.add(Flatten())\n",
    "#model.add(Dense(512))\n",
    "model.add(Dense(64))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(nb_classes))\n",
    "model.add(Activation('softmax'))\n",
    "\n",
    "# let's train the model using SGD + momentum (how original).\n",
    "sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "Xtrain = Xtrain_norm.astype('float32')\n",
    "Xtest = Xtest_norm.astype('float32')\n",
    "\n",
    "if not data_augmentation:\n",
    "    print('Not using data augmentation.')\n",
    "    model.fit(Xtrain_norm, Ytrain,\n",
    "              batch_size=batch_size,\n",
    "              nb_epoch=nb_epoch,\n",
    "              validation_data=(Xtest_norm, Ytest),\n",
    "              shuffle=True)\n",
    "else:\n",
    "    print('Using real-time data augmentation.')\n",
    "\n",
    "    # this will do preprocessing and realtime data augmentation\n",
    "    datagen = ImageDataGenerator(\n",
    "        featurewise_center=False,  # set input mean to 0 over the dataset\n",
    "        samplewise_center=False,  # set each sample mean to 0\n",
    "        featurewise_std_normalization=False,  # divide inputs by std of the dataset\n",
    "        samplewise_std_normalization=False,  # divide each input by its std\n",
    "        zca_whitening=False,  # apply ZCA whitening\n",
    "        rotation_range=0,  # randomly rotate images in the range (degrees, 0 to 180)\n",
    "        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)\n",
    "        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)\n",
    "        horizontal_flip=True,  # randomly flip images\n",
    "        vertical_flip=False)  # randomly flip images\n",
    "\n",
    "    # compute quantities required for featurewise normalization\n",
    "    # (std, mean, and principal components if ZCA whitening is applied)\n",
    "    datagen.fit(Xtrain_norm)\n",
    "\n",
    "    # fit the model on the batches generated by datagen.flow()\n",
    "    model.fit_generator(datagen.flow(Xtrain_norm, Ytrain,\n",
    "                        batch_size=batch_size),\n",
    "                        samples_per_epoch=Xtrain_norm.shape[0],\n",
    "                        nb_epoch=nb_epoch,\n",
    "                        validation_data=(Xtest_norm, Ytest))\n",
    "    \n",
    "    model.save('new_net')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We only trained on 1 epoch due to the limitation on the time for the neural network to run (about 40 minutest per epoch). Looks like the network is about 78% accurate. Below we'll notice that the model is predicing 1's for all the samples in the test set. That can't be good. \n",
    "\n",
    "Since we did a roughly 80/20 split of the data, it seems as if the guessing all ones is giving the best predictive power. i.e. it looks like we are underfitting (?) so that the model can't really tell the difference between the two categories. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30336/30336 [==============================] - 203s    \n"
     ]
    }
   ],
   "source": [
    "#look at samples that were classified incorrectly, and correctly, and plot them. \n",
    "classes = model.predict_classes(Xtest, batch_size=2048)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30336/30336 [==============================] - 202s    \n"
     ]
    }
   ],
   "source": [
    "proba = model.predict_proba(Xtest, batch_size=2048)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-1, 2)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJYAAAJPCAYAAAAudl6uAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAHOFJREFUeJzt3W+MZXd93/HP12xJS2gR0LIgG9sUSCnIkoPUrSO3ylQo\nwTgNphHlTyO1pWqFKG4ipVJNUiS2z+BBJUIRQk6dFKMgQyLxLzGNqeg0oirGBVwcsGFRZWNcs0Hl\nTwXkgQPfPphre7zMrM33npk7u369pCufe+5vzjlX+vnsznvPPbe6OwAAAADw47pg0wcAAAAAwLlJ\nWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYGTtsFRVF1XVJ6rqC1V1R1X9yj7j3lFV\np6rq9qq6fN39AgAAALBZxxbYxp8n+bXuvr2qnpzkM1V1S3ff9eCAqnpZkud29/Or6m8neXeSKxbY\nNwAAAAAbsvYVS9399e6+fbX83SR3JrnwjGHXJLlxNebWJE+pquPr7hsAAACAzVn0HktVdWmSy5Pc\nesZLFya5d9fz+/Kj8QkAAACAc8hiYWn1MbjfT/KrqyuXAAAAADiPLXGPpVTVsexEpfd294f3GHJf\nkmfven7Rat1e2+oljgkAAACAh3V3Lb3NRcJSkt9O8sXu/s19Xv9IkjcmeX9VXZHk2919er+NdWtL\nHL6TJ0/m5MmTmz4MHofMPTbJ/GNTzD02yfxjU8w9Nqlq8aaUZIGwVFVXJvnlJHdU1eeSdJLfSHJJ\nku7u67v75qq6uqq+kuR7SV637n4BAAAA2Ky1w1J3//ckT3gM465dd18AAAAAHB2LfiscnMu2trY2\nfQg8Tpl7bJL5x6aYe2yS+cemmHucj+qo3c+oqvqoHRMAAADAuayqDuTm3a5YAgAAAGBEWAIAAABg\nRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBE\nWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERY\nAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgC\nAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIA\nAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAA\nAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAA\nYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABg\nRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBE\nWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERY\nAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBEWAIAAABgRFgCAAAAYERYAgAAAGBkkbBUVTdU1emq\n+vw+r/9sVX27qj67erx5if0CAAAAsDnHFtrO7yT5D0luPMuYP+7uly+0PwAAAAA2bJErlrr7k0m+\n9SjDaol9AQAAAHA0HOY9ln6mqm6vqj+sqhce4n4BAAAAOABLfRTu0XwmycXd/f2qelmSDyX5qUPa\nNwAAAAAH4FDCUnd/d9fyx6rqXVX1tO7+5l7jT548+dDy1tZWtra2DvwYAQAAAM4X29vb2d7ePvD9\nVHcvs6GqS5N8tLsv2+O14919erV8IskHuvvSfbbTSx0TAAAAAElVpbsXv//1IlcsVdX7kmwleXpV\nfTXJW5I8MUl39/VJXllVb0jyQJI/S/LqJfYLAAAAwOYsdsXSUlyxBAAAALCsg7pi6TC/FQ4AAACA\n84iwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADA\niLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCI\nsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiw\nBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAE\nAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQA\nAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAA\nAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAA\nwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADA\niLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCI\nsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiw\nBAAAAMCIsAQAAADAyCJhqapuqKrTVfX5s4x5R1Wdqqrbq+ryJfYLAAAAwOYsdcXS7yR56X4vVtXL\nkjy3u5+f5PVJ3r3QfgEAAADYkEXCUnd/Msm3zjLkmiQ3rsbemuQpVXV8iX0DAAAAsBmHdY+lC5Pc\nu+v5fat1AAAAAJyjjm36APbyi7+46SMAAACAzXnTm5Irr9z0UcCjO6ywdF+SZ+96ftFq3Z6e/OST\nDy1fdtlWLrts66COCwAAAI6c5zxn00fAuW57ezvb29sHvp/q7mU2VHVpko9292V7vHZ1kjd29y9U\n1RVJ3t7dV+yznV7qmAAAAABIqirdXUtvd5ErlqrqfUm2kjy9qr6a5C1Jnpiku/v67r65qq6uqq8k\n+V6S1y2xXwAAAAA2Z7ErlpbiiiUAAACAZR3UFUuH9a1wAAAAAJxnhCUAAAAARoQlAAAAAEaEJQAA\nAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAA\nAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAA\nRoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABG\nhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaE\nJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQl\nAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUA\nAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAA\nAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAA\nAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAA\nRoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEYWCUtVdVVV\n3VVVX66q6/Z4/Wer6ttV9dnV481L7BcAAACAzTm27gaq6oIk70zykiT/J8ltVfXh7r7rjKF/3N0v\nX3d/AAAAABwNS1yxdCLJqe6+p7sfSHJTkmv2GFcL7AsAAACAI2KJsHRhknt3Pf/aat2Zfqaqbq+q\nP6yqFy6wXwAAAAA2aO2Pwj1Gn0lycXd/v6peluRDSX7qkPYNAAAAwAFYIizdl+TiXc8vWq17SHd/\nd9fyx6rqXVX1tO7+5l4bPHny5EPLW1tb2draWuAwAQAAAB4ftre3s729feD7qe5ebwNVT0jypezc\nvPv+JJ9O8truvnPXmOPdfXq1fCLJB7r70n221+seEwAAAAAPq6p09+L3v177iqXu/kFVXZvkluzc\ns+mG7r6zql6/83Jfn+SVVfWGJA8k+bMkr153vwAAAABs1tpXLC3NFUsAAAAAyzqoK5aW+FY4AAAA\nAB6HhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAA\nAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAA\nRoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABG\nhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaE\nJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQl\nAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUA\nAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAA\nAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAA\nAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAA\nRoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABG\nhCUAAAAARoQlAAAAAEYWCUtVdVVV3VVVX66q6/YZ846qOlVVt1fV5UvsFwAAAIDNWTssVdUFSd6Z\n5KVJXpTktVX1gjPGvCzJc7v7+Ulen+Td6+4XAAAAgM1a4oqlE0lOdfc93f1AkpuSXHPGmGuS3Jgk\n3X1rkqdU1fEF9g0AAADAhiwRli5Mcu+u519brTvbmPv2GAMAAADAOcTNuwEAAAAYObbANu5LcvGu\n5xet1p055tmPMuYhJ0+efGh5a2srW1tb6x4jAAAAwOPG9vZ2tre3D3w/1d3rbaDqCUm+lOQlSe5P\n8ukkr+3uO3eNuTrJG7v7F6rqiiRv7+4r9tler3tMAAAAADysqtLdtfR2175iqbt/UFXXJrklOx+t\nu6G776yq1++83Nd3981VdXVVfSXJ95K8bt39AgAAALBZa1+xtDRXLAEAAAAs66CuWHLzbgAAAABG\nhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaE\nJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQl\nAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUA\nAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAA\nAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAA\nAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAA\nRoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABG\nhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaE\nJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQl\nAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUAAAAARoQlAAAAAEaEJQAAAABGhCUA\nAAAARoQlAAAAAEaOrfPDVfXUJO9PckmSu5O8qru/s8e4u5N8J8kPkzzQ3SfW2S8AAAAAm7fuFUtv\nSvJfuvtvJPlEkl/fZ9wPk2x190+LSgAAAADnh3XD0jVJ3rNafk+SV+wzrhbYFwAAAABHyLqx5xnd\nfTpJuvvrSZ6xz7hO8vGquq2q/sWa+wQAAADgCHjUeyxV1ceTHN+9Kjuh6M17DO99NnNld99fVX8t\nO4Hpzu7+5I99tAAAAAAcGY8alrr75/Z7rapOV9Xx7j5dVc9M8qf7bOP+1X+/UVUfTHIiyb5h6eTJ\nkw8tb21tZWtr69EOEwAAAICV7e3tbG9vH/h+qnu/i4weww9XvS3JN7v7bVV1XZKndvebzhjzpCQX\ndPd3q+onk9yS5N919y37bLPXOSYAAAAAHqmq0t21+HbXDEtPS/KBJM9Ock+SV3X3t6vqWUl+q7v/\nflU9J8kHs/MxuWNJfre733qWbQpLAAAAAAs6kmHpIAhLAAAAAMs6qLC07rfCAQAAAPA4JSwBAAAA\nMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAw\nIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAi\nLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIs\nAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwB\nAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEA\nAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAA\nADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAA\nMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAw\nIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAi\nLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIsAQAAADAiLAEAAAAwIiwBAAAAMCIs\nAQAAADCyVliqqldW1Z9U1Q+q6sVnGXdVVd1VVV+uquvW2ScAAAAAR8O6VyzdkeQfJPlv+w2oqguS\nvDPJS5O8KMlrq+oFa+4XFre9vb3pQ+Bxytxjk8w/NsXcY5PMPzbF3ON8tFZY6u4vdfepJHWWYSeS\nnOrue7r7gSQ3Jblmnf3CQXCSZ1PMPTbJ/GNTzD02yfxjU8w9zkeHcY+lC5Pcu+v511brAAAAADiH\nHXu0AVX18STHd69K0kn+bXd/9KAODAAAAICjrbp7/Y1U/dck/7q7P7vHa1ckOdndV62evylJd/fb\n9tnW+gcEAAAAwCN099luZTTyqFcs/Rj2O7jbkjyvqi5Jcn+S1yR57X4bOYg3CQAAAMDy1rrHUlW9\noqruTXJFkj+oqo+t1j+rqv4gSbr7B0muTXJLki8kuam771zvsAEAAADYtEU+CgcAAADA489hfCvc\nY1JVV1XVXVX15aq6btPHw/mhqu6uqv9VVZ+rqk+v1j21qm6pqi9V1R9V1VN2jf/1qjpVVXdW1c/v\nWv/iqvr8an6+fRPvhaOvqm6oqtNV9fld6xabb1X1xKq6afUz/6OqLj68d8dRts/ce0tVfa2qPrt6\nXLXrNXOPRVTVRVX1iar6QlXdUVW/slrv3MeB22P+/avVeuc/DlRV/URV3br6HeOOqnrLar1zHwfq\nLHNvs+e97t74IzuB6ytJLknyF5LcnuQFmz4uj3P/keR/J3nqGeveluTfrJavS/LW1fILk3wuO/ce\nu3Q1Jx+8qu/WJH9rtXxzkpdu+r15HL1Hkr+T5PIkn9+1brH5luQNSd61Wn51dj5avPH37bH5xz5z\n7y1Jfm2PsX/T3PNY6pHkmUkuXy0/OcmXkrzAuc/jMB5nmX/Ofx4H/kjypNV/n5DkU0lOOPd5HMZj\nn7m30fPeUbli6USSU919T3c/kOSmJNds+Jg4P1R+9Mq8a5K8Z7X8niSvWC2/PDv/0/x5d9+d5FSS\nE1X1zCR/ubtvW427cdfPwEO6+5NJvnXG6iXn2+5t/X6Slyz+Jjgn7TP3kr2/WOOamHsspLu/3t23\nr5a/m+TOJBfFuY9DsM/8u3D1svMfB6q7v79a/Ins/NLece7jEOwz95INnveOSli6MMm9u55/LQ//\noQDr6CQfr6rbquqfr9Yd7+7Tyc5fSJI8Y7X+zHl432rdhdmZkw8yP/lxPGPB+fbQz/TOFyN8u6qe\ndnCHznng2qq6var+467L8c09DkRVXZqdK+c+lWX/rDX/eFS75t+tq1XOfxyoqrqgqj6X5OtJPr76\nBd25jwO3z9xLNnjeOyphCQ7Kld394iRXJ3ljVf3dPFx0H+QO9hymJefbXv8qAQ96V5K/3t2XZ+cv\nHv9+wW2bezxCVT05O/+q+aurK0cO8s9a849H2GP+Of9x4Lr7h93909m5SvNEVb0ozn0cgj3m3guz\n4fPeUQlL9yXZfUOoi1brYC3dff/qv99I8qHsfOzydFUdT5LVJYB/uhp+X5Jn7/rxB+fhfuvhsVhy\nvj30WlU9Iclf6e5vHtyhcy7r7m/06sPxSX4rO+e/xNxjYVV1LDu/1L+3uz+8Wu3cx6HYa/45/3GY\nuvv/JdlOclWc+zhEu+feps97RyUs3ZbkeVV1SVU9Mclrknxkw8fEOa6qnrT6F6xU1U8m+fkkd2Rn\nbv3T1bB/kuTBvwR/JMlrVnfBf06S5yX59Ooy1u9U1YmqqiT/eNfPwJkqj6z6S863j6y2kST/MMkn\nDuxdcC56xNxb/YX2Qb+U5E9Wy+YeS/vtJF/s7t/ctc65j8PyI/PP+Y+DVlV/9cGPGlXVX0ryc9m5\nx5dzHwdqn7l316bPe8fWfF+L6O4fVNW1SW7JTuy6obvv3PBhce47nuSDVdXZmeu/2923VNX/TPKB\nqvpnSe5J8qok6e4vVtUHknwxyQNJ/uWu6vvGJP8pyV9McnN3/+fDfSucC6rqfUm2kjy9qr6anW9n\neGuS31tovt2Q5L1VdSrJ/81OhIf95t7fq6rLk/wwyd1JXp+Yeyyrqq5M8stJ7ljd76GT/EZ2vhlp\nqT9rzT/2dJb594+c/zhgz0rynqq6IDu/v76/u2+uqk/FuY+Dtd/cu3GT5716eJsAAAAA8NgdlY/C\nAQAAAHCOEZYAAAAAGBGWAAAAABgRlgAAAAAYEZYAAAAAGBGWAAAAABgRlgAAAAAYEZYAAAAAGPn/\nygyEjinjcW4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2ac4fd3bb650>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#save the model.\n",
    "#run on separate data.\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.plot(classes)\n",
    "plt.ylim(-1,2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[    0  6569]\n",
      " [    0 23767]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x2ac51d36b2d0>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARQAAAD0CAYAAABaQ9lPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njQOKmZXGAcXMSuOAYmal+f9psN9jiVdsJwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2ac51d2ef990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(ytest,classes)\n",
    "print cm\n",
    "\n",
    "plt.matshow(cm)\n",
    "plt.title('Confusion Matrix')\n",
    "plt.colorbar()\n",
    "plt.xlabel('Predicted Label')\n",
    "plt.ylabel('True Label')"
   ]
  },
  {
   "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
}