Product centricity- NN(2).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "import keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def csv_to_numpy_array(filePath, delimiter):\n",
    "    return np.genfromtxt(filePath, delimiter=delimiter, dtype=float,skip_header=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### read input files "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "trainX = csv_to_numpy_array(\"../prod_centricity/trainX.csv\", delimiter=\",\")\n",
    "trainY = csv_to_numpy_array(\"../prod_centricity/trainY.csv\", delimiter=\",\")\n",
    "testX = csv_to_numpy_array(\"../prod_centricity/testX.csv\", delimiter=\",\")\n",
    "testY = csv_to_numpy_array(\"../prod_centricity/testY.csv\", delimiter=\",\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### preprocess x to normalize input (DONT do on labels) for faster convergence and to avoid numerical stability issues like overflow or underflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.preprocessing import normalize\n",
    "from sklearn import preprocessing\n",
    "trainX_norm = preprocessing.scale(trainX)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### initialize the model with fully connected layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Build the Fully Connected Neural Network in Keras Here\n",
    "model = Sequential()\n",
    "model.add(Dense(10,input_dim=4,name='layer1'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(10,name='layer2'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dropout(0.3))\n",
    "model.add(Dense(2))\n",
    "model.add(Activation('softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/300\n",
      "211/211 [==============================] - 0s - loss: 0.6360 - acc: 0.6635     \n",
      "Epoch 2/300\n",
      "211/211 [==============================] - 0s - loss: 0.6093 - acc: 0.6588     \n",
      "Epoch 3/300\n",
      "211/211 [==============================] - 0s - loss: 0.6032 - acc: 0.6493     \n",
      "Epoch 4/300\n",
      "211/211 [==============================] - 0s - loss: 0.5738 - acc: 0.7488     \n",
      "Epoch 5/300\n",
      "211/211 [==============================] - 0s - loss: 0.5859 - acc: 0.6682     \n",
      "Epoch 6/300\n",
      "211/211 [==============================] - 0s - loss: 0.5566 - acc: 0.7393     \n",
      "Epoch 7/300\n",
      "211/211 [==============================] - 0s - loss: 0.5339 - acc: 0.7109     \n",
      "Epoch 8/300\n",
      "211/211 [==============================] - 0s - loss: 0.5085 - acc: 0.7725     \n",
      "Epoch 9/300\n",
      "211/211 [==============================] - 0s - loss: 0.5412 - acc: 0.7251     \n",
      "Epoch 10/300\n",
      "211/211 [==============================] - 0s - loss: 0.5250 - acc: 0.7346     \n",
      "Epoch 11/300\n",
      "211/211 [==============================] - 0s - loss: 0.4879 - acc: 0.7678     \n",
      "Epoch 12/300\n",
      "211/211 [==============================] - 0s - loss: 0.4949 - acc: 0.7536     \n",
      "Epoch 13/300\n",
      "211/211 [==============================] - 0s - loss: 0.4962 - acc: 0.7725     \n",
      "Epoch 14/300\n",
      "211/211 [==============================] - 0s - loss: 0.4836 - acc: 0.7915     \n",
      "Epoch 15/300\n",
      "211/211 [==============================] - 0s - loss: 0.4709 - acc: 0.7820     \n",
      "Epoch 16/300\n",
      "211/211 [==============================] - 0s - loss: 0.4621 - acc: 0.8294     \n",
      "Epoch 17/300\n",
      "211/211 [==============================] - 0s - loss: 0.4817 - acc: 0.7867     \n",
      "Epoch 18/300\n",
      "211/211 [==============================] - 0s - loss: 0.4379 - acc: 0.8389     \n",
      "Epoch 19/300\n",
      "211/211 [==============================] - 0s - loss: 0.4493 - acc: 0.8152     \n",
      "Epoch 20/300\n",
      "211/211 [==============================] - 0s - loss: 0.4679 - acc: 0.8057     \n",
      "Epoch 21/300\n",
      "211/211 [==============================] - 0s - loss: 0.4238 - acc: 0.8483     \n",
      "Epoch 22/300\n",
      "211/211 [==============================] - 0s - loss: 0.4261 - acc: 0.8152     \n",
      "Epoch 23/300\n",
      "211/211 [==============================] - 0s - loss: 0.4585 - acc: 0.8057     \n",
      "Epoch 24/300\n",
      "211/211 [==============================] - 0s - loss: 0.4147 - acc: 0.8199     \n",
      "Epoch 25/300\n",
      "211/211 [==============================] - 0s - loss: 0.4219 - acc: 0.8389     \n",
      "Epoch 26/300\n",
      "211/211 [==============================] - 0s - loss: 0.4303 - acc: 0.8246     \n",
      "Epoch 27/300\n",
      "211/211 [==============================] - 0s - loss: 0.3978 - acc: 0.8341     \n",
      "Epoch 28/300\n",
      "211/211 [==============================] - 0s - loss: 0.4153 - acc: 0.8294     \n",
      "Epoch 29/300\n",
      "211/211 [==============================] - 0s - loss: 0.4448 - acc: 0.8246     \n",
      "Epoch 30/300\n",
      "211/211 [==============================] - 0s - loss: 0.4289 - acc: 0.8531     \n",
      "Epoch 31/300\n",
      "211/211 [==============================] - 0s - loss: 0.4248 - acc: 0.8578     \n",
      "Epoch 32/300\n",
      "211/211 [==============================] - 0s - loss: 0.3871 - acc: 0.8531     \n",
      "Epoch 33/300\n",
      "211/211 [==============================] - 0s - loss: 0.3910 - acc: 0.8436     \n",
      "Epoch 34/300\n",
      "211/211 [==============================] - 0s - loss: 0.4023 - acc: 0.8246     \n",
      "Epoch 35/300\n",
      "211/211 [==============================] - 0s - loss: 0.4062 - acc: 0.8720     \n",
      "Epoch 36/300\n",
      "211/211 [==============================] - 0s - loss: 0.4240 - acc: 0.8246     \n",
      "Epoch 37/300\n",
      "211/211 [==============================] - 0s - loss: 0.3992 - acc: 0.8341     \n",
      "Epoch 38/300\n",
      "211/211 [==============================] - 0s - loss: 0.3771 - acc: 0.8626     \n",
      "Epoch 39/300\n",
      "211/211 [==============================] - 0s - loss: 0.3869 - acc: 0.8531     \n",
      "Epoch 40/300\n",
      "211/211 [==============================] - 0s - loss: 0.4204 - acc: 0.8389     \n",
      "Epoch 41/300\n",
      "211/211 [==============================] - 0s - loss: 0.3937 - acc: 0.8341     \n",
      "Epoch 42/300\n",
      "211/211 [==============================] - 0s - loss: 0.4287 - acc: 0.8673     \n",
      "Epoch 43/300\n",
      "211/211 [==============================] - 0s - loss: 0.4148 - acc: 0.8389     \n",
      "Epoch 44/300\n",
      "211/211 [==============================] - 0s - loss: 0.3949 - acc: 0.8626     \n",
      "Epoch 45/300\n",
      "211/211 [==============================] - 0s - loss: 0.4042 - acc: 0.8531     \n",
      "Epoch 46/300\n",
      "211/211 [==============================] - 0s - loss: 0.4036 - acc: 0.8389     \n",
      "Epoch 47/300\n",
      "211/211 [==============================] - 0s - loss: 0.3791 - acc: 0.8720     \n",
      "Epoch 48/300\n",
      "211/211 [==============================] - 0s - loss: 0.3605 - acc: 0.8626     \n",
      "Epoch 49/300\n",
      "211/211 [==============================] - 0s - loss: 0.3929 - acc: 0.8578     \n",
      "Epoch 50/300\n",
      "211/211 [==============================] - 0s - loss: 0.4013 - acc: 0.8436     \n",
      "Epoch 51/300\n",
      "211/211 [==============================] - 0s - loss: 0.3514 - acc: 0.8483     \n",
      "Epoch 52/300\n",
      "211/211 [==============================] - 0s - loss: 0.3827 - acc: 0.8720     \n",
      "Epoch 53/300\n",
      "211/211 [==============================] - 0s - loss: 0.3425 - acc: 0.8768     \n",
      "Epoch 54/300\n",
      "211/211 [==============================] - 0s - loss: 0.3562 - acc: 0.8578     \n",
      "Epoch 55/300\n",
      "211/211 [==============================] - 0s - loss: 0.3805 - acc: 0.8531     \n",
      "Epoch 56/300\n",
      "211/211 [==============================] - 0s - loss: 0.3841 - acc: 0.8578     \n",
      "Epoch 57/300\n",
      "211/211 [==============================] - 0s - loss: 0.3872 - acc: 0.8578     \n",
      "Epoch 58/300\n",
      "211/211 [==============================] - 0s - loss: 0.3329 - acc: 0.8720     \n",
      "Epoch 59/300\n",
      "211/211 [==============================] - 0s - loss: 0.3921 - acc: 0.8768     \n",
      "Epoch 60/300\n",
      "211/211 [==============================] - 0s - loss: 0.3618 - acc: 0.8768     \n",
      "Epoch 61/300\n",
      "211/211 [==============================] - 0s - loss: 0.3870 - acc: 0.8531     \n",
      "Epoch 62/300\n",
      "211/211 [==============================] - 0s - loss: 0.3895 - acc: 0.8578     \n",
      "Epoch 63/300\n",
      "211/211 [==============================] - 0s - loss: 0.3790 - acc: 0.8720     \n",
      "Epoch 64/300\n",
      "211/211 [==============================] - 0s - loss: 0.3453 - acc: 0.8626     \n",
      "Epoch 65/300\n",
      "211/211 [==============================] - 0s - loss: 0.3777 - acc: 0.8436     \n",
      "Epoch 66/300\n",
      "211/211 [==============================] - 0s - loss: 0.4007 - acc: 0.8720     \n",
      "Epoch 67/300\n",
      "211/211 [==============================] - 0s - loss: 0.3860 - acc: 0.8626     \n",
      "Epoch 68/300\n",
      "211/211 [==============================] - 0s - loss: 0.3853 - acc: 0.8389     \n",
      "Epoch 69/300\n",
      "211/211 [==============================] - 0s - loss: 0.3650 - acc: 0.8626     \n",
      "Epoch 70/300\n",
      "211/211 [==============================] - 0s - loss: 0.3799 - acc: 0.8531     \n",
      "Epoch 71/300\n",
      "211/211 [==============================] - 0s - loss: 0.3759 - acc: 0.8673     \n",
      "Epoch 72/300\n",
      "211/211 [==============================] - 0s - loss: 0.3632 - acc: 0.8626     \n",
      "Epoch 73/300\n",
      "211/211 [==============================] - 0s - loss: 0.3489 - acc: 0.8436     \n",
      "Epoch 74/300\n",
      "211/211 [==============================] - 0s - loss: 0.3912 - acc: 0.8389     \n",
      "Epoch 75/300\n",
      "211/211 [==============================] - 0s - loss: 0.3623 - acc: 0.8578     \n",
      "Epoch 76/300\n",
      "211/211 [==============================] - 0s - loss: 0.4176 - acc: 0.8578     \n",
      "Epoch 77/300\n",
      "211/211 [==============================] - 0s - loss: 0.3849 - acc: 0.8578     \n",
      "Epoch 78/300\n",
      "211/211 [==============================] - 0s - loss: 0.3464 - acc: 0.8815     \n",
      "Epoch 79/300\n",
      "211/211 [==============================] - 0s - loss: 0.3591 - acc: 0.8815     \n",
      "Epoch 80/300\n",
      "211/211 [==============================] - 0s - loss: 0.3691 - acc: 0.8436     \n",
      "Epoch 81/300\n",
      "211/211 [==============================] - 0s - loss: 0.3373 - acc: 0.8863     \n",
      "Epoch 82/300\n",
      "211/211 [==============================] - 0s - loss: 0.3744 - acc: 0.8483     \n",
      "Epoch 83/300\n",
      "211/211 [==============================] - 0s - loss: 0.3522 - acc: 0.8815     \n",
      "Epoch 84/300\n",
      "211/211 [==============================] - 0s - loss: 0.3789 - acc: 0.8483     \n",
      "Epoch 85/300\n",
      "211/211 [==============================] - 0s - loss: 0.3940 - acc: 0.8531     \n",
      "Epoch 86/300\n",
      "211/211 [==============================] - 0s - loss: 0.3685 - acc: 0.9005     \n",
      "Epoch 87/300\n",
      "211/211 [==============================] - 0s - loss: 0.3438 - acc: 0.8673     \n",
      "Epoch 88/300\n",
      "211/211 [==============================] - 0s - loss: 0.3909 - acc: 0.8531     \n",
      "Epoch 89/300\n",
      "211/211 [==============================] - 0s - loss: 0.3600 - acc: 0.8578     \n",
      "Epoch 90/300\n",
      "211/211 [==============================] - 0s - loss: 0.3501 - acc: 0.8863     \n",
      "Epoch 91/300\n",
      "211/211 [==============================] - 0s - loss: 0.3772 - acc: 0.8578     \n",
      "Epoch 92/300\n",
      "211/211 [==============================] - 0s - loss: 0.4001 - acc: 0.8626     \n",
      "Epoch 93/300\n",
      "211/211 [==============================] - 0s - loss: 0.3660 - acc: 0.8531     \n",
      "Epoch 94/300\n",
      "211/211 [==============================] - 0s - loss: 0.3578 - acc: 0.8673     \n",
      "Epoch 95/300\n",
      "211/211 [==============================] - 0s - loss: 0.3776 - acc: 0.8578     \n",
      "Epoch 96/300\n",
      "211/211 [==============================] - 0s - loss: 0.3891 - acc: 0.8673     \n",
      "Epoch 97/300\n",
      "211/211 [==============================] - 0s - loss: 0.3594 - acc: 0.8673     \n",
      "Epoch 98/300\n",
      "211/211 [==============================] - 0s - loss: 0.3651 - acc: 0.8341     \n",
      "Epoch 99/300\n",
      "211/211 [==============================] - 0s - loss: 0.3692 - acc: 0.8436     \n",
      "Epoch 100/300\n",
      "211/211 [==============================] - 0s - loss: 0.3768 - acc: 0.8720     \n",
      "Epoch 101/300\n",
      "211/211 [==============================] - 0s - loss: 0.3448 - acc: 0.8863     \n",
      "Epoch 102/300\n",
      "211/211 [==============================] - 0s - loss: 0.3482 - acc: 0.8768     \n",
      "Epoch 103/300\n",
      "211/211 [==============================] - 0s - loss: 0.3514 - acc: 0.8483     \n",
      "Epoch 104/300\n",
      "211/211 [==============================] - 0s - loss: 0.3849 - acc: 0.8389     \n",
      "Epoch 105/300\n",
      "211/211 [==============================] - 0s - loss: 0.3705 - acc: 0.8720     \n",
      "Epoch 106/300\n",
      "211/211 [==============================] - 0s - loss: 0.3505 - acc: 0.8768     \n",
      "Epoch 107/300\n",
      "211/211 [==============================] - 0s - loss: 0.3257 - acc: 0.8957     \n",
      "Epoch 108/300\n",
      "211/211 [==============================] - 0s - loss: 0.3526 - acc: 0.8578     \n",
      "Epoch 109/300\n",
      "211/211 [==============================] - 0s - loss: 0.3753 - acc: 0.8910     \n",
      "Epoch 110/300\n",
      "211/211 [==============================] - 0s - loss: 0.3659 - acc: 0.8815     \n",
      "Epoch 111/300\n",
      "211/211 [==============================] - 0s - loss: 0.3273 - acc: 0.8578     \n",
      "Epoch 112/300\n",
      "211/211 [==============================] - 0s - loss: 0.3501 - acc: 0.8815     \n",
      "Epoch 113/300\n",
      "211/211 [==============================] - 0s - loss: 0.3488 - acc: 0.8768     \n",
      "Epoch 114/300\n",
      "211/211 [==============================] - 0s - loss: 0.3321 - acc: 0.8673     \n",
      "Epoch 115/300\n",
      "211/211 [==============================] - 0s - loss: 0.3520 - acc: 0.8720     \n",
      "Epoch 116/300\n",
      "211/211 [==============================] - 0s - loss: 0.3063 - acc: 0.8863     \n",
      "Epoch 117/300\n",
      "211/211 [==============================] - 0s - loss: 0.3455 - acc: 0.8863     \n",
      "Epoch 118/300\n",
      "211/211 [==============================] - 0s - loss: 0.3421 - acc: 0.8531     \n",
      "Epoch 119/300\n",
      "211/211 [==============================] - 0s - loss: 0.3807 - acc: 0.8626     \n",
      "Epoch 120/300\n",
      "211/211 [==============================] - 0s - loss: 0.4093 - acc: 0.8720     \n",
      "Epoch 121/300\n",
      "211/211 [==============================] - 0s - loss: 0.3349 - acc: 0.8910     \n",
      "Epoch 122/300\n",
      "211/211 [==============================] - 0s - loss: 0.3864 - acc: 0.8626     \n",
      "Epoch 123/300\n",
      "211/211 [==============================] - 0s - loss: 0.3498 - acc: 0.8910     \n",
      "Epoch 124/300\n",
      "211/211 [==============================] - 0s - loss: 0.3492 - acc: 0.8673     \n",
      "Epoch 125/300\n",
      "211/211 [==============================] - 0s - loss: 0.3192 - acc: 0.8815     \n",
      "Epoch 126/300\n",
      "211/211 [==============================] - 0s - loss: 0.3377 - acc: 0.8720     \n",
      "Epoch 127/300\n",
      "211/211 [==============================] - 0s - loss: 0.3730 - acc: 0.8720     \n",
      "Epoch 128/300\n",
      "211/211 [==============================] - 0s - loss: 0.3810 - acc: 0.8341     \n",
      "Epoch 129/300\n",
      "211/211 [==============================] - 0s - loss: 0.3416 - acc: 0.8815     \n",
      "Epoch 130/300\n",
      "211/211 [==============================] - 0s - loss: 0.3577 - acc: 0.8626     \n",
      "Epoch 131/300\n",
      "211/211 [==============================] - 0s - loss: 0.3530 - acc: 0.8389     \n",
      "Epoch 132/300\n",
      "211/211 [==============================] - 0s - loss: 0.3640 - acc: 0.8626     \n",
      "Epoch 133/300\n",
      "211/211 [==============================] - 0s - loss: 0.3613 - acc: 0.8578     \n",
      "Epoch 134/300\n",
      "211/211 [==============================] - 0s - loss: 0.3485 - acc: 0.8626     \n",
      "Epoch 135/300\n",
      "211/211 [==============================] - 0s - loss: 0.3619 - acc: 0.8673     \n",
      "Epoch 136/300\n",
      "211/211 [==============================] - 0s - loss: 0.3719 - acc: 0.8531     \n",
      "Epoch 137/300\n",
      "211/211 [==============================] - 0s - loss: 0.3688 - acc: 0.8768     \n",
      "Epoch 138/300\n",
      "211/211 [==============================] - 0s - loss: 0.3537 - acc: 0.8673     \n",
      "Epoch 139/300\n",
      "211/211 [==============================] - 0s - loss: 0.3570 - acc: 0.8578     \n",
      "Epoch 140/300\n",
      "211/211 [==============================] - 0s - loss: 0.3502 - acc: 0.8483     \n",
      "Epoch 141/300\n",
      "211/211 [==============================] - 0s - loss: 0.3255 - acc: 0.8720     \n",
      "Epoch 142/300\n",
      "211/211 [==============================] - 0s - loss: 0.3477 - acc: 0.8768     \n",
      "Epoch 143/300\n",
      "211/211 [==============================] - 0s - loss: 0.3319 - acc: 0.8768     \n",
      "Epoch 144/300\n",
      "211/211 [==============================] - 0s - loss: 0.3614 - acc: 0.8531     \n",
      "Epoch 145/300\n",
      "211/211 [==============================] - 0s - loss: 0.3219 - acc: 0.8815     \n",
      "Epoch 146/300\n",
      "211/211 [==============================] - 0s - loss: 0.3549 - acc: 0.8815     \n",
      "Epoch 147/300\n",
      "211/211 [==============================] - 0s - loss: 0.3491 - acc: 0.8531     \n",
      "Epoch 148/300\n",
      "211/211 [==============================] - 0s - loss: 0.3651 - acc: 0.8768     \n",
      "Epoch 149/300\n",
      "211/211 [==============================] - 0s - loss: 0.3699 - acc: 0.8626     \n",
      "Epoch 150/300\n",
      "211/211 [==============================] - 0s - loss: 0.3526 - acc: 0.8768     \n",
      "Epoch 151/300\n",
      "211/211 [==============================] - 0s - loss: 0.3658 - acc: 0.8720     \n",
      "Epoch 152/300\n",
      "211/211 [==============================] - 0s - loss: 0.3595 - acc: 0.8626     \n",
      "Epoch 153/300\n",
      "211/211 [==============================] - 0s - loss: 0.3565 - acc: 0.8626     \n",
      "Epoch 154/300\n",
      "211/211 [==============================] - 0s - loss: 0.3669 - acc: 0.8578     \n",
      "Epoch 155/300\n",
      "211/211 [==============================] - 0s - loss: 0.3530 - acc: 0.8720     \n",
      "Epoch 156/300\n",
      "211/211 [==============================] - 0s - loss: 0.3824 - acc: 0.8673     \n",
      "Epoch 157/300\n",
      "211/211 [==============================] - 0s - loss: 0.3430 - acc: 0.8626     \n",
      "Epoch 158/300\n",
      "211/211 [==============================] - 0s - loss: 0.3732 - acc: 0.8626     \n",
      "Epoch 159/300\n",
      "211/211 [==============================] - 0s - loss: 0.3338 - acc: 0.8768     \n",
      "Epoch 160/300\n",
      "211/211 [==============================] - 0s - loss: 0.3258 - acc: 0.8815     \n",
      "Epoch 161/300\n",
      "211/211 [==============================] - 0s - loss: 0.3641 - acc: 0.8436     \n",
      "Epoch 162/300\n",
      "211/211 [==============================] - 0s - loss: 0.3578 - acc: 0.8626     \n",
      "Epoch 163/300\n",
      "211/211 [==============================] - 0s - loss: 0.3697 - acc: 0.8673     \n",
      "Epoch 164/300\n",
      "211/211 [==============================] - 0s - loss: 0.3544 - acc: 0.8768     \n",
      "Epoch 165/300\n",
      "211/211 [==============================] - 0s - loss: 0.3451 - acc: 0.8768     \n",
      "Epoch 166/300\n",
      "211/211 [==============================] - 0s - loss: 0.3532 - acc: 0.8626     \n",
      "Epoch 167/300\n",
      "211/211 [==============================] - 0s - loss: 0.3606 - acc: 0.8483     \n",
      "Epoch 168/300\n",
      "211/211 [==============================] - 0s - loss: 0.3767 - acc: 0.8673     \n",
      "Epoch 169/300\n",
      "211/211 [==============================] - 0s - loss: 0.3532 - acc: 0.8626     \n",
      "Epoch 170/300\n",
      "211/211 [==============================] - 0s - loss: 0.3413 - acc: 0.8863     \n",
      "Epoch 171/300\n",
      "211/211 [==============================] - 0s - loss: 0.4144 - acc: 0.8578     \n",
      "Epoch 172/300\n",
      "211/211 [==============================] - 0s - loss: 0.3329 - acc: 0.8720     \n",
      "Epoch 173/300\n",
      "211/211 [==============================] - 0s - loss: 0.3516 - acc: 0.8578     \n",
      "Epoch 174/300\n",
      "211/211 [==============================] - 0s - loss: 0.3816 - acc: 0.8768     \n",
      "Epoch 175/300\n",
      "211/211 [==============================] - 0s - loss: 0.3512 - acc: 0.8626     \n",
      "Epoch 176/300\n",
      "211/211 [==============================] - 0s - loss: 0.3274 - acc: 0.8673     \n",
      "Epoch 177/300\n",
      "211/211 [==============================] - 0s - loss: 0.3258 - acc: 0.8626     \n",
      "Epoch 178/300\n",
      "211/211 [==============================] - 0s - loss: 0.3460 - acc: 0.8673     \n",
      "Epoch 179/300\n",
      "211/211 [==============================] - 0s - loss: 0.3560 - acc: 0.8815     \n",
      "Epoch 180/300\n",
      "211/211 [==============================] - 0s - loss: 0.3254 - acc: 0.8768     \n",
      "Epoch 181/300\n",
      "211/211 [==============================] - 0s - loss: 0.3463 - acc: 0.8626     \n",
      "Epoch 182/300\n",
      "211/211 [==============================] - 0s - loss: 0.3706 - acc: 0.8531     \n",
      "Epoch 183/300\n",
      "211/211 [==============================] - 0s - loss: 0.3227 - acc: 0.8673     \n",
      "Epoch 184/300\n",
      "211/211 [==============================] - 0s - loss: 0.3245 - acc: 0.8720     \n",
      "Epoch 185/300\n",
      "211/211 [==============================] - 0s - loss: 0.3409 - acc: 0.8626     \n",
      "Epoch 186/300\n",
      "211/211 [==============================] - 0s - loss: 0.3554 - acc: 0.8815     \n",
      "Epoch 187/300\n",
      "211/211 [==============================] - 0s - loss: 0.3278 - acc: 0.8768     \n",
      "Epoch 188/300\n",
      "211/211 [==============================] - 0s - loss: 0.3530 - acc: 0.8720     \n",
      "Epoch 189/300\n",
      "211/211 [==============================] - 0s - loss: 0.3531 - acc: 0.8436     \n",
      "Epoch 190/300\n",
      "211/211 [==============================] - 0s - loss: 0.3745 - acc: 0.8768     \n",
      "Epoch 191/300\n",
      "211/211 [==============================] - 0s - loss: 0.3201 - acc: 0.8815     \n",
      "Epoch 192/300\n",
      "211/211 [==============================] - 0s - loss: 0.3423 - acc: 0.8578     \n",
      "Epoch 193/300\n",
      "211/211 [==============================] - 0s - loss: 0.3640 - acc: 0.8673     \n",
      "Epoch 194/300\n",
      "211/211 [==============================] - 0s - loss: 0.3256 - acc: 0.8863     \n",
      "Epoch 195/300\n",
      "211/211 [==============================] - 0s - loss: 0.3532 - acc: 0.8768     \n",
      "Epoch 196/300\n",
      "211/211 [==============================] - 0s - loss: 0.3605 - acc: 0.8768     \n",
      "Epoch 197/300\n",
      "211/211 [==============================] - 0s - loss: 0.3511 - acc: 0.8626     \n",
      "Epoch 198/300\n",
      "211/211 [==============================] - 0s - loss: 0.3510 - acc: 0.8815     \n",
      "Epoch 199/300\n",
      "211/211 [==============================] - 0s - loss: 0.3208 - acc: 0.8815     \n",
      "Epoch 200/300\n",
      "211/211 [==============================] - 0s - loss: 0.3084 - acc: 0.8768     \n",
      "Epoch 201/300\n",
      "211/211 [==============================] - 0s - loss: 0.3427 - acc: 0.8815     \n",
      "Epoch 202/300\n",
      "211/211 [==============================] - 0s - loss: 0.3400 - acc: 0.8768     \n",
      "Epoch 203/300\n",
      "211/211 [==============================] - 0s - loss: 0.3624 - acc: 0.8626     \n",
      "Epoch 204/300\n",
      "211/211 [==============================] - 0s - loss: 0.3942 - acc: 0.8626     \n",
      "Epoch 205/300\n",
      "211/211 [==============================] - 0s - loss: 0.3693 - acc: 0.8578     \n",
      "Epoch 206/300\n",
      "211/211 [==============================] - 0s - loss: 0.3230 - acc: 0.8436     \n",
      "Epoch 207/300\n",
      "211/211 [==============================] - 0s - loss: 0.3368 - acc: 0.8720     \n",
      "Epoch 208/300\n",
      "211/211 [==============================] - 0s - loss: 0.3485 - acc: 0.8578     \n",
      "Epoch 209/300\n",
      "211/211 [==============================] - 0s - loss: 0.3415 - acc: 0.8673     \n",
      "Epoch 210/300\n",
      "211/211 [==============================] - 0s - loss: 0.3437 - acc: 0.8815     \n",
      "Epoch 211/300\n",
      "211/211 [==============================] - 0s - loss: 0.3295 - acc: 0.8815     \n",
      "Epoch 212/300\n",
      "211/211 [==============================] - 0s - loss: 0.3229 - acc: 0.8910     \n",
      "Epoch 213/300\n",
      "211/211 [==============================] - 0s - loss: 0.3422 - acc: 0.8673     \n",
      "Epoch 214/300\n",
      "211/211 [==============================] - 0s - loss: 0.3422 - acc: 0.8673     \n",
      "Epoch 215/300\n",
      "211/211 [==============================] - 0s - loss: 0.3666 - acc: 0.8720     \n",
      "Epoch 216/300\n",
      "211/211 [==============================] - 0s - loss: 0.3382 - acc: 0.8626     \n",
      "Epoch 217/300\n",
      "211/211 [==============================] - 0s - loss: 0.3429 - acc: 0.8673     \n",
      "Epoch 218/300\n",
      "211/211 [==============================] - 0s - loss: 0.3547 - acc: 0.8673     \n",
      "Epoch 219/300\n",
      "211/211 [==============================] - 0s - loss: 0.3314 - acc: 0.8673     \n",
      "Epoch 220/300\n",
      "211/211 [==============================] - 0s - loss: 0.3359 - acc: 0.8673     \n",
      "Epoch 221/300\n",
      "211/211 [==============================] - 0s - loss: 0.3388 - acc: 0.8673     \n",
      "Epoch 222/300\n",
      "211/211 [==============================] - 0s - loss: 0.3456 - acc: 0.8389     \n",
      "Epoch 223/300\n",
      "211/211 [==============================] - 0s - loss: 0.3458 - acc: 0.8673     \n",
      "Epoch 224/300\n",
      "211/211 [==============================] - 0s - loss: 0.3482 - acc: 0.8815     \n",
      "Epoch 225/300\n",
      "211/211 [==============================] - 0s - loss: 0.3331 - acc: 0.8483     \n",
      "Epoch 226/300\n",
      "211/211 [==============================] - 0s - loss: 0.3438 - acc: 0.8910     \n",
      "Epoch 227/300\n",
      "211/211 [==============================] - 0s - loss: 0.3290 - acc: 0.8863     \n",
      "Epoch 228/300\n",
      "211/211 [==============================] - 0s - loss: 0.3524 - acc: 0.8626     \n",
      "Epoch 229/300\n",
      "211/211 [==============================] - 0s - loss: 0.3265 - acc: 0.8673     \n",
      "Epoch 230/300\n",
      "211/211 [==============================] - 0s - loss: 0.3543 - acc: 0.8768     \n",
      "Epoch 231/300\n",
      "211/211 [==============================] - 0s - loss: 0.3475 - acc: 0.8720     \n",
      "Epoch 232/300\n",
      "211/211 [==============================] - 0s - loss: 0.3583 - acc: 0.8673     \n",
      "Epoch 233/300\n",
      "211/211 [==============================] - 0s - loss: 0.3616 - acc: 0.8815     \n",
      "Epoch 234/300\n",
      "211/211 [==============================] - 0s - loss: 0.3499 - acc: 0.8626     \n",
      "Epoch 235/300\n",
      "211/211 [==============================] - 0s - loss: 0.3415 - acc: 0.8720     \n",
      "Epoch 236/300\n",
      "211/211 [==============================] - 0s - loss: 0.3522 - acc: 0.8673     \n",
      "Epoch 237/300\n",
      "211/211 [==============================] - 0s - loss: 0.3218 - acc: 0.8910     \n",
      "Epoch 238/300\n",
      "211/211 [==============================] - 0s - loss: 0.3806 - acc: 0.8626     \n",
      "Epoch 239/300\n",
      "211/211 [==============================] - 0s - loss: 0.3363 - acc: 0.8957     \n",
      "Epoch 240/300\n",
      "211/211 [==============================] - 0s - loss: 0.3585 - acc: 0.8626     \n",
      "Epoch 241/300\n",
      "211/211 [==============================] - 0s - loss: 0.3263 - acc: 0.8768     \n",
      "Epoch 242/300\n",
      "211/211 [==============================] - 0s - loss: 0.3591 - acc: 0.8863     \n",
      "Epoch 243/300\n",
      "211/211 [==============================] - 0s - loss: 0.3476 - acc: 0.8720     \n",
      "Epoch 244/300\n",
      "211/211 [==============================] - 0s - loss: 0.3271 - acc: 0.8720     \n",
      "Epoch 245/300\n",
      "211/211 [==============================] - 0s - loss: 0.3489 - acc: 0.8720     \n",
      "Epoch 246/300\n",
      "211/211 [==============================] - 0s - loss: 0.3471 - acc: 0.8768     \n",
      "Epoch 247/300\n",
      "211/211 [==============================] - 0s - loss: 0.3439 - acc: 0.8673     \n",
      "Epoch 248/300\n",
      "211/211 [==============================] - 0s - loss: 0.3559 - acc: 0.8673     \n",
      "Epoch 249/300\n",
      "211/211 [==============================] - 0s - loss: 0.3685 - acc: 0.8626     \n",
      "Epoch 250/300\n",
      "211/211 [==============================] - 0s - loss: 0.3635 - acc: 0.8626     \n",
      "Epoch 251/300\n",
      "211/211 [==============================] - 0s - loss: 0.3267 - acc: 0.8815     \n",
      "Epoch 252/300\n",
      "211/211 [==============================] - 0s - loss: 0.3276 - acc: 0.8768     \n",
      "Epoch 253/300\n",
      "211/211 [==============================] - 0s - loss: 0.3559 - acc: 0.8863     \n",
      "Epoch 254/300\n",
      "211/211 [==============================] - 0s - loss: 0.3305 - acc: 0.8768     \n",
      "Epoch 255/300\n",
      "211/211 [==============================] - 0s - loss: 0.3335 - acc: 0.8863     \n",
      "Epoch 256/300\n",
      "211/211 [==============================] - 0s - loss: 0.3491 - acc: 0.8483     \n",
      "Epoch 257/300\n",
      "211/211 [==============================] - 0s - loss: 0.3040 - acc: 0.8768     \n",
      "Epoch 258/300\n",
      "211/211 [==============================] - 0s - loss: 0.3253 - acc: 0.8720     \n",
      "Epoch 259/300\n",
      "211/211 [==============================] - 0s - loss: 0.3263 - acc: 0.8768     \n",
      "Epoch 260/300\n",
      "211/211 [==============================] - 0s - loss: 0.3497 - acc: 0.8673     \n",
      "Epoch 261/300\n",
      "211/211 [==============================] - 0s - loss: 0.3399 - acc: 0.8720     \n",
      "Epoch 262/300\n",
      "211/211 [==============================] - 0s - loss: 0.3473 - acc: 0.8720     \n",
      "Epoch 263/300\n",
      "211/211 [==============================] - 0s - loss: 0.3292 - acc: 0.8578     \n",
      "Epoch 264/300\n",
      "211/211 [==============================] - 0s - loss: 0.3279 - acc: 0.8863     \n",
      "Epoch 265/300\n",
      "211/211 [==============================] - 0s - loss: 0.3629 - acc: 0.8863     \n",
      "Epoch 266/300\n",
      "211/211 [==============================] - 0s - loss: 0.3085 - acc: 0.8768     \n",
      "Epoch 267/300\n",
      "211/211 [==============================] - 0s - loss: 0.3417 - acc: 0.8531     \n",
      "Epoch 268/300\n",
      "211/211 [==============================] - 0s - loss: 0.3279 - acc: 0.8768     \n",
      "Epoch 269/300\n",
      "211/211 [==============================] - 0s - loss: 0.3088 - acc: 0.8815     \n",
      "Epoch 270/300\n",
      "211/211 [==============================] - 0s - loss: 0.3158 - acc: 0.8957     \n",
      "Epoch 271/300\n",
      "211/211 [==============================] - 0s - loss: 0.3586 - acc: 0.8531     \n",
      "Epoch 272/300\n",
      "211/211 [==============================] - 0s - loss: 0.3377 - acc: 0.8957     \n",
      "Epoch 273/300\n",
      "211/211 [==============================] - 0s - loss: 0.3741 - acc: 0.8768     \n",
      "Epoch 274/300\n",
      "211/211 [==============================] - 0s - loss: 0.3337 - acc: 0.8815     \n",
      "Epoch 275/300\n",
      "211/211 [==============================] - 0s - loss: 0.3306 - acc: 0.8673     \n",
      "Epoch 276/300\n",
      "211/211 [==============================] - 0s - loss: 0.3337 - acc: 0.8863     \n",
      "Epoch 277/300\n",
      "211/211 [==============================] - 0s - loss: 0.3506 - acc: 0.8815     \n",
      "Epoch 278/300\n",
      "211/211 [==============================] - 0s - loss: 0.3572 - acc: 0.8483     \n",
      "Epoch 279/300\n",
      "211/211 [==============================] - 0s - loss: 0.3331 - acc: 0.8578     \n",
      "Epoch 280/300\n",
      "211/211 [==============================] - 0s - loss: 0.3308 - acc: 0.8626     \n",
      "Epoch 281/300\n",
      "211/211 [==============================] - 0s - loss: 0.3582 - acc: 0.8531     \n",
      "Epoch 282/300\n",
      "211/211 [==============================] - 0s - loss: 0.3197 - acc: 0.8815     \n",
      "Epoch 283/300\n",
      "211/211 [==============================] - 0s - loss: 0.3186 - acc: 0.8815     \n",
      "Epoch 284/300\n",
      "211/211 [==============================] - 0s - loss: 0.3557 - acc: 0.8768     \n",
      "Epoch 285/300\n",
      "211/211 [==============================] - 0s - loss: 0.3317 - acc: 0.8578     \n",
      "Epoch 286/300\n",
      "211/211 [==============================] - 0s - loss: 0.3598 - acc: 0.8673     \n",
      "Epoch 287/300\n",
      "211/211 [==============================] - 0s - loss: 0.3490 - acc: 0.8673     \n",
      "Epoch 288/300\n",
      "211/211 [==============================] - 0s - loss: 0.3231 - acc: 0.8910     \n",
      "Epoch 289/300\n",
      "211/211 [==============================] - 0s - loss: 0.3545 - acc: 0.8863     \n",
      "Epoch 290/300\n",
      "211/211 [==============================] - 0s - loss: 0.3343 - acc: 0.8626     \n",
      "Epoch 291/300\n",
      "211/211 [==============================] - 0s - loss: 0.3254 - acc: 0.8483     \n",
      "Epoch 292/300\n",
      "211/211 [==============================] - 0s - loss: 0.3082 - acc: 0.8673     \n",
      "Epoch 293/300\n",
      "211/211 [==============================] - 0s - loss: 0.3177 - acc: 0.8957     \n",
      "Epoch 294/300\n",
      "211/211 [==============================] - 0s - loss: 0.3230 - acc: 0.8673     \n",
      "Epoch 295/300\n",
      "211/211 [==============================] - 0s - loss: 0.3033 - acc: 0.8768     \n",
      "Epoch 296/300\n",
      "211/211 [==============================] - 0s - loss: 0.2964 - acc: 0.8863     \n",
      "Epoch 297/300\n",
      "211/211 [==============================] - 0s - loss: 0.3193 - acc: 0.8815     \n",
      "Epoch 298/300\n",
      "211/211 [==============================] - 0s - loss: 0.3277 - acc: 0.8768     \n",
      "Epoch 299/300\n",
      "211/211 [==============================] - 0s - loss: 0.3564 - acc: 0.8578     \n",
      "Epoch 300/300\n",
      "211/211 [==============================] - 0s - loss: 0.3456 - acc: 0.8720     \n"
     ]
    }
   ],
   "source": [
    "from keras.callbacks import TensorBoard\n",
    "model.compile('adam', 'binary_crossentropy', ['accuracy'])\n",
    "history = model.fit(trainX_norm, trainY, nb_epoch=300,callbacks=[TensorBoard(log_dir='./Graph',histogram_freq=0,write_graph=True)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### lets evaluate on our test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('Test score:', 0.3503632925029071)\n",
      "('Test accuracy:', 0.8867924528301887)\n"
     ]
    }
   ],
   "source": [
    "score = model.evaluate(preprocessing.scale(testX), testY, verbose=0)\n",
    "print('Test score:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### lets visualize our model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: Could not load \"/home/local/ANT/badeepak/anaconda2/lib/graphviz/libgvplugin_pango.so.6\" - file not found\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/svg+xml": [
       "<svg height=\"751pt\" viewBox=\"0.00 0.00 300.55 751.40\" width=\"301pt\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g class=\"graph\" id=\"graph0\" transform=\"scale(1 1) rotate(0) translate(4 747.4)\">\n",
       "<title>G</title>\n",
       "<polygon fill=\"white\" points=\"-4,4 -4,-747.4 296.553,-747.4 296.553,4 -4,4\" stroke=\"none\"/>\n",
       "<!-- 139718602484112 -->\n",
       "<g class=\"node\" id=\"node1\"><title>139718602484112</title>\n",
       "<polygon fill=\"none\" points=\"0,-693.3 0,-742.9 292.553,-742.9 292.553,-693.3 0,-693.3\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"83.6196\" y=\"-713.9\">dense_input_7: InputLayer</text>\n",
       "<polyline fill=\"none\" points=\"167.239,-693.3 167.239,-742.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"195.075\" y=\"-726.3\">input:</text>\n",
       "<polyline fill=\"none\" points=\"167.239,-718.1 222.911,-718.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"195.075\" y=\"-701.5\">output:</text>\n",
       "<polyline fill=\"none\" points=\"222.911,-693.3 222.911,-742.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"257.732\" y=\"-726.3\">(None, 4)</text>\n",
       "<polyline fill=\"none\" points=\"222.911,-718.1 292.553,-718.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"257.732\" y=\"-701.5\">(None, 4)</text>\n",
       "</g>\n",
       "<!-- 139718602483920 -->\n",
       "<g class=\"node\" id=\"node2\"><title>139718602483920</title>\n",
       "<polygon fill=\"none\" points=\"33.4439,-606.7 33.4439,-656.3 259.11,-656.3 259.11,-606.7 33.4439,-606.7\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-627.3\">layer1: Dense</text>\n",
       "<polyline fill=\"none\" points=\"126.795,-606.7 126.795,-656.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"154.631\" y=\"-639.7\">input:</text>\n",
       "<polyline fill=\"none\" points=\"126.795,-631.5 182.467,-631.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"154.631\" y=\"-614.9\">output:</text>\n",
       "<polyline fill=\"none\" points=\"182.467,-606.7 182.467,-656.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"220.288\" y=\"-639.7\">(None, 4)</text>\n",
       "<polyline fill=\"none\" points=\"182.467,-631.5 259.109,-631.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"220.788\" y=\"-614.9\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718602484112&#45;&gt;139718602483920 -->\n",
       "<g class=\"edge\" id=\"edge1\"><title>139718602484112-&gt;139718602483920</title>\n",
       "<path d=\"M146.277,-693.083C146.277,-684.867 146.277,-675.512 146.277,-666.652\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-666.577 146.277,-656.577 142.777,-666.577 149.777,-666.577\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718602727312 -->\n",
       "<g class=\"node\" id=\"node3\"><title>139718602727312</title>\n",
       "<polygon fill=\"none\" points=\"0.7721,-520.1 0.7721,-569.7 291.781,-569.7 291.781,-520.1 0.7721,-520.1\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-540.7\">activation_16: Activation</text>\n",
       "<polyline fill=\"none\" points=\"159.467,-520.1 159.467,-569.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"187.303\" y=\"-553.1\">input:</text>\n",
       "<polyline fill=\"none\" points=\"159.467,-544.9 215.139,-544.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"187.303\" y=\"-528.3\">output:</text>\n",
       "<polyline fill=\"none\" points=\"215.139,-520.1 215.139,-569.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-553.1\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"215.139,-544.9 291.781,-544.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-528.3\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718602483920&#45;&gt;139718602727312 -->\n",
       "<g class=\"edge\" id=\"edge2\"><title>139718602483920-&gt;139718602727312</title>\n",
       "<path d=\"M146.277,-606.483C146.277,-598.267 146.277,-588.912 146.277,-580.052\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-579.977 146.277,-569.977 142.777,-579.977 149.777,-579.977\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718795617168 -->\n",
       "<g class=\"node\" id=\"node4\"><title>139718795617168</title>\n",
       "<polygon fill=\"none\" points=\"16.3198,-433.5 16.3198,-483.1 276.234,-483.1 276.234,-433.5 16.3198,-433.5\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-454.1\">dropout_3: Dropout</text>\n",
       "<polyline fill=\"none\" points=\"143.919,-433.5 143.919,-483.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"171.755\" y=\"-466.5\">input:</text>\n",
       "<polyline fill=\"none\" points=\"143.919,-458.3 199.591,-458.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"171.755\" y=\"-441.7\">output:</text>\n",
       "<polyline fill=\"none\" points=\"199.591,-433.5 199.591,-483.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"237.912\" y=\"-466.5\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"199.591,-458.3 276.234,-458.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"237.912\" y=\"-441.7\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718602727312&#45;&gt;139718795617168 -->\n",
       "<g class=\"edge\" id=\"edge3\"><title>139718602727312-&gt;139718795617168</title>\n",
       "<path d=\"M146.277,-519.883C146.277,-511.667 146.277,-502.312 146.277,-493.452\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-493.377 146.277,-483.377 142.777,-493.377 149.777,-493.377\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718601550224 -->\n",
       "<g class=\"node\" id=\"node5\"><title>139718601550224</title>\n",
       "<polygon fill=\"none\" points=\"33.4439,-346.9 33.4439,-396.5 259.11,-396.5 259.11,-346.9 33.4439,-346.9\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-367.5\">layer2: Dense</text>\n",
       "<polyline fill=\"none\" points=\"126.795,-346.9 126.795,-396.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"154.631\" y=\"-379.9\">input:</text>\n",
       "<polyline fill=\"none\" points=\"126.795,-371.7 182.467,-371.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"154.631\" y=\"-355.1\">output:</text>\n",
       "<polyline fill=\"none\" points=\"182.467,-346.9 182.467,-396.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"220.788\" y=\"-379.9\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"182.467,-371.7 259.109,-371.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"220.788\" y=\"-355.1\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718795617168&#45;&gt;139718601550224 -->\n",
       "<g class=\"edge\" id=\"edge4\"><title>139718795617168-&gt;139718601550224</title>\n",
       "<path d=\"M146.277,-433.283C146.277,-425.067 146.277,-415.712 146.277,-406.852\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-406.777 146.277,-396.777 142.777,-406.777 149.777,-406.777\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718606848528 -->\n",
       "<g class=\"node\" id=\"node6\"><title>139718606848528</title>\n",
       "<polygon fill=\"none\" points=\"0.7721,-260.3 0.7721,-309.9 291.781,-309.9 291.781,-260.3 0.7721,-260.3\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-280.9\">activation_17: Activation</text>\n",
       "<polyline fill=\"none\" points=\"159.467,-260.3 159.467,-309.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"187.303\" y=\"-293.3\">input:</text>\n",
       "<polyline fill=\"none\" points=\"159.467,-285.1 215.139,-285.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"187.303\" y=\"-268.5\">output:</text>\n",
       "<polyline fill=\"none\" points=\"215.139,-260.3 215.139,-309.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-293.3\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"215.139,-285.1 291.781,-285.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-268.5\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718601550224&#45;&gt;139718606848528 -->\n",
       "<g class=\"edge\" id=\"edge5\"><title>139718601550224-&gt;139718606848528</title>\n",
       "<path d=\"M146.277,-346.683C146.277,-338.467 146.277,-329.112 146.277,-320.252\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-320.177 146.277,-310.177 142.777,-320.177 149.777,-320.177\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718602775184 -->\n",
       "<g class=\"node\" id=\"node7\"><title>139718602775184</title>\n",
       "<polygon fill=\"none\" points=\"16.3198,-173.7 16.3198,-223.3 276.234,-223.3 276.234,-173.7 16.3198,-173.7\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-194.3\">dropout_4: Dropout</text>\n",
       "<polyline fill=\"none\" points=\"143.919,-173.7 143.919,-223.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"171.755\" y=\"-206.7\">input:</text>\n",
       "<polyline fill=\"none\" points=\"143.919,-198.5 199.591,-198.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"171.755\" y=\"-181.9\">output:</text>\n",
       "<polyline fill=\"none\" points=\"199.591,-173.7 199.591,-223.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"237.912\" y=\"-206.7\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"199.591,-198.5 276.234,-198.5 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"237.912\" y=\"-181.9\">(None, 10)</text>\n",
       "</g>\n",
       "<!-- 139718606848528&#45;&gt;139718602775184 -->\n",
       "<g class=\"edge\" id=\"edge6\"><title>139718606848528-&gt;139718602775184</title>\n",
       "<path d=\"M146.277,-260.083C146.277,-251.867 146.277,-242.512 146.277,-233.652\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-233.577 146.277,-223.577 142.777,-233.577 149.777,-233.577\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718601760720 -->\n",
       "<g class=\"node\" id=\"node8\"><title>139718601760720</title>\n",
       "<polygon fill=\"none\" points=\"24.4972,-87.1 24.4972,-136.7 268.056,-136.7 268.056,-87.1 24.4972,-87.1\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"80.1196\" y=\"-107.7\">dense_19: Dense</text>\n",
       "<polyline fill=\"none\" points=\"135.742,-87.1 135.742,-136.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"163.578\" y=\"-120.1\">input:</text>\n",
       "<polyline fill=\"none\" points=\"135.742,-111.9 191.414,-111.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"163.578\" y=\"-95.3\">output:</text>\n",
       "<polyline fill=\"none\" points=\"191.414,-87.1 191.414,-136.7 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"229.735\" y=\"-120.1\">(None, 10)</text>\n",
       "<polyline fill=\"none\" points=\"191.414,-111.9 268.056,-111.9 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"229.235\" y=\"-95.3\">(None, 2)</text>\n",
       "</g>\n",
       "<!-- 139718602775184&#45;&gt;139718601760720 -->\n",
       "<g class=\"edge\" id=\"edge7\"><title>139718602775184-&gt;139718601760720</title>\n",
       "<path d=\"M146.277,-173.483C146.277,-165.267 146.277,-155.912 146.277,-147.052\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-146.977 146.277,-136.977 142.777,-146.977 149.777,-146.977\" stroke=\"black\"/>\n",
       "</g>\n",
       "<!-- 139718605430544 -->\n",
       "<g class=\"node\" id=\"node9\"><title>139718605430544</title>\n",
       "<polygon fill=\"none\" points=\"4.2721,-0.5 4.2721,-50.1 288.281,-50.1 288.281,-0.5 4.2721,-0.5\" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"83.6196\" y=\"-21.1\">activation_18: Activation</text>\n",
       "<polyline fill=\"none\" points=\"162.967,-0.5 162.967,-50.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"190.803\" y=\"-33.5\">input:</text>\n",
       "<polyline fill=\"none\" points=\"162.967,-25.3 218.639,-25.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"190.803\" y=\"-8.7\">output:</text>\n",
       "<polyline fill=\"none\" points=\"218.639,-0.5 218.639,-50.1 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-33.5\">(None, 2)</text>\n",
       "<polyline fill=\"none\" points=\"218.639,-25.3 288.281,-25.3 \" stroke=\"black\"/>\n",
       "<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"253.46\" y=\"-8.7\">(None, 2)</text>\n",
       "</g>\n",
       "<!-- 139718601760720&#45;&gt;139718605430544 -->\n",
       "<g class=\"edge\" id=\"edge8\"><title>139718601760720-&gt;139718605430544</title>\n",
       "<path d=\"M146.277,-86.8833C146.277,-78.6673 146.277,-69.3117 146.277,-60.4518\" fill=\"none\" stroke=\"black\"/>\n",
       "<polygon fill=\"black\" points=\"149.777,-60.3767 146.277,-50.3767 142.777,-60.3768 149.777,-60.3767\" stroke=\"black\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>"
      ],
      "text/plain": [
       "<IPython.core.display.SVG object>"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import SVG\n",
    "from keras.utils.visualize_util import model_to_dot\n",
    "SVG(model_to_dot(model,show_shapes = True,show_layer_names = True).create(prog='dot', format='svg'))\n",
    "#from keras.utils.visualize_util import plot\n",
    "#plot(model, to_file='nn_prodcentricity_model.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "keras.callbacks.History"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tbcallback = keras.callbacks.TensorBoard(log_dir='./Graph', histogram_freq=0,  write_graph=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/local/ANT/badeepak/imagesnaps'"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "os.getcwd()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tensorboard --logdir imagesnaps/Graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['acc', 'loss']"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "step=range(1,301)\n",
    "acc=history.history['acc']\n",
    "loss=history.history['loss']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### lets visualize accuracy and loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f875b56f4d0>]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbAAAAEaCAYAAABwyQKiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFkIIoSb9Pz6r4XCPgQ+TAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87f02c4610>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "fig = plt.figure()\n",
    "# Create two subplots on their own axes and give titles\n",
    "ax1 = plt.subplot(\"211\")\n",
    "ax1.set_title(\"TRAINING ACCURACY\", fontsize=18)\n",
    "ax2 = plt.subplot(\"212\")\n",
    "ax2.set_title(\"TRAINING COST\", fontsize=18)\n",
    "plt.tight_layout()\n",
    "\n",
    "ax1.plot(step, acc)\n",
    "ax2.plot(step, loss)\n",
    "#fig.canvas.draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f8759df6210>]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2AAAAGqCAYAAABtfidfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfc3CXwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f875a1ce690>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "fig = plt.figure(figsize=(12, 6))\n",
    "# Create two subplots on their own axes and give titles\n",
    "ax1 = plt.subplot(\"221\")\n",
    "ax1.set_title(\"TRAINING ACCURACY\", fontsize=18)\n",
    "ax2 = plt.subplot(\"222\")\n",
    "ax2.set_title(\"TRAINING COST\", fontsize=18)\n",
    "ax3 = plt.subplot(\"223\")\n",
    "ax3.set_title(\"TRAINING ACCURACY(with dropout)\", fontsize=18)\n",
    "ax4 = plt.subplot(\"224\")\n",
    "ax4.set_title(\"TRAINING COST(with dropout)\", fontsize=18)\n",
    "plt.tight_layout()\n",
    "\n",
    "\n",
    "ax1.plot(step, acc)\n",
    "ax2.plot(step, loss)\n",
    "ax3.plot(step, acc1)\n",
    "ax4.plot(step, loss1)\n",
    "#fig.canvas.draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "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
}