Digit_recog_MLP_jpy_notebook .ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Welcome to Jupyter!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can also get this tutorial and run it on your laptop:\n",
    "\n",
    "    git clone /~https://github.com/ipython/ipython-in-depth\n",
    "\n",
    "Install IPython and Jupyter:\n",
    "\n",
    "with [conda](https://www.anaconda.com/download):\n",
    "\n",
    "    conda install ipython jupyter\n",
    "\n",
    "with pip:\n",
    "\n",
    "    # first, always upgrade pip!\n",
    "    pip install --upgrade pip\n",
    "    pip install --upgrade ipython jupyter\n",
    "\n",
    "Start the notebook in the tutorial directory:\n",
    "\n",
    "    cd ipython-in-depth\n",
    "    jupyter notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Hand-written Digit Recognition\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Throught this following notebook we shall be using a package called scikit-learn(sklearn).<br>\n",
    "This module contains all the requried resources we'll be needing for our digit-recognizer.<br>\n",
    "The following items are imported from the mod:\n",
    "   * datasets = contains the data requried for our program(data from images of hand-written digits)\n",
    "   * svm = learning model \n",
    "   * cross_validation.train_test_split = (helper funtion)which we shall use to split out dataset into training and testing samples.\n",
    "   * matplotlib.pyplot = for plotting out the images "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip3 install sklearn\n",
    "#!pip3 install matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Importing dependencies\n",
    "from sklearn import datasets,svm\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn.neural_network as nn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "#IMPORT DATA\n",
    "dataset=datasets.load_digits()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optical Recognition of Handwritten Digits Data Set\n",
      "===================================================\n",
      "\n",
      "Notes\n",
      "-----\n",
      "Data Set Characteristics:\n",
      "    :Number of Instances: 5620\n",
      "    :Number of Attributes: 64\n",
      "    :Attribute Information: 8x8 image of integer pixels in the range 0..16.\n",
      "    :Missing Attribute Values: None\n",
      "    :Creator: E. Alpaydin (alpaydin '@' boun.edu.tr)\n",
      "    :Date: July; 1998\n",
      "\n",
      "This is a copy of the test set of the UCI ML hand-written digits datasets\n",
      "http://archive.ics.uci.edu/ml/datasets/Optical+Recognition+of+Handwritten+Digits\n",
      "\n",
      "The data set contains images of hand-written digits: 10 classes where\n",
      "each class refers to a digit.\n",
      "\n",
      "Preprocessing programs made available by NIST were used to extract\n",
      "normalized bitmaps of handwritten digits from a preprinted form. From a\n",
      "total of 43 people, 30 contributed to the training set and different 13\n",
      "to the test set. 32x32 bitmaps are divided into nonoverlapping blocks of\n",
      "4x4 and the number of on pixels are counted in each block. This generates\n",
      "an input matrix of 8x8 where each element is an integer in the range\n",
      "0..16. This reduces dimensionality and gives invariance to small\n",
      "distortions.\n",
      "\n",
      "For info on NIST preprocessing routines, see M. D. Garris, J. L. Blue, G.\n",
      "T. Candela, D. L. Dimmick, J. Geist, P. J. Grother, S. A. Janet, and C.\n",
      "L. Wilson, NIST Form-Based Handprint Recognition System, NISTIR 5469,\n",
      "1994.\n",
      "\n",
      "References\n",
      "----------\n",
      "  - C. Kaynak (1995) Methods of Combining Multiple Classifiers and Their\n",
      "    Applications to Handwritten Digit Recognition, MSc Thesis, Institute of\n",
      "    Graduate Studies in Science and Engineering, Bogazici University.\n",
      "  - E. Alpaydin, C. Kaynak (1998) Cascading Classifiers, Kybernetika.\n",
      "  - Ken Tang and Ponnuthurai N. Suganthan and Xi Yao and A. Kai Qin.\n",
      "    Linear dimensionalityreduction using relevance weighted LDA. School of\n",
      "    Electrical and Electronic Engineering Nanyang Technological University.\n",
      "    2005.\n",
      "  - Claudio Gentile. A New Approximate Maximal Margin Classification\n",
      "    Algorithm. NIPS. 2000.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(dataset.DESCR)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "* Let us store our images and labels of what those images are in variables img_data and output_labels respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "img_data=dataset.images\n",
    "output_label=dataset.target"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Lets have a look at some image data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output Label :  0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output Label :  1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output Label :  2\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Display few images with labels\n",
    "for i in range(3):\n",
    "    plt.imshow(img_data[i])\n",
    "    plt.show()\n",
    "    print(\"Output Label : \",output_label[i])\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## How do we make the machine understand our data ?\n",
    "\n",
    "Our training data is in the form of images.But how does the computer interpret it?\n",
    "The best was for us to represent our data is in the form of vectors.\n",
    "Here we have 2-D , 8x8 images and we need to reshape each image it into one vector(i.e- of lenght 64)<br>\n",
    "This step is often reffered to as <b>data-preprocessing</b>.<br>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output Label :  0\n",
      "The vector i is : \n",
      " [ 0.  0.  5. 13.  9.  1.  0.  0.  0.  0. 13. 15. 10. 15.  5.  0.  0.  3.\n",
      " 15.  2.  0. 11.  8.  0.  0.  4. 12.  0.  0.  8.  8.  0.  0.  5.  8.  0.\n",
      "  0.  9.  8.  0.  0.  4. 11.  0.  1. 12.  7.  0.  0.  2. 14.  5. 10. 12.\n",
      "  0.  0.  0.  0.  6. 13. 10.  0.  0.  0.]\n",
      "Shape of i : (1, 64)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "'''\n",
    "RESHAPE IMAGES(FLATTEN)\n",
    "We can show how the image is converted to a column vector. \n",
    "Original Img-dim = 8x8\n",
    "Result vector = 1x64 \n",
    "'''\n",
    "#Taking one sample image as i \n",
    "i=img_data[0]\n",
    "\n",
    "#Check what the image is \n",
    "plt.imshow(img_data[0])\n",
    "plt.show()\n",
    "print (\"Output Label : \",output_label[0])\n",
    "\n",
    "#Vectorizing the 2-D image\n",
    "i=i.reshape((64,1))\n",
    "#Transposing the column vector to be stored in the dataset\n",
    "i=i.T \n",
    "\n",
    "#Showing i is same as training_data\n",
    "print (\"The vector i is : \\n\",i[0])\n",
    "print(\"Shape of i :\",i.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lucky for us!\n",
    "\n",
    "The digits dataset we have ,already has the images converted into a numpy array.We will be using this numerical data now. :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#And so now we shall use the dataset.data for furthur use. \n",
    "training_data=dataset.data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.  0.  5. 13.  9.  1.  0.  0.  0.  0. 13. 15. 10. 15.  5.  0.  0.  3.\n",
      " 15.  2.  0. 11.  8.  0.  0.  4. 12.  0.  0.  8.  8.  0.  0.  5.  8.  0.\n",
      "  0.  9.  8.  0.  0.  4. 11.  0.  1. 12.  7.  0.  0.  2. 14.  5. 10. 12.\n",
      "  0.  0.  0.  0.  6. 13. 10.  0.  0.  0.]\n",
      "Shape :  (1797, 64)\n"
     ]
    }
   ],
   "source": [
    "#Lets make sure its the same \n",
    "print(training_data[0])\n",
    "print(\"Shape : \",training_data.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of samples :  1797\n",
      "Number of training samples :  1347\n",
      "Number of testing samples :  450\n"
     ]
    }
   ],
   "source": [
    "#TRAIN-TEST SPLIT\n",
    "\n",
    "X_train,X_test,y_train,y_test = train_test_split(training_data,output_label)\n",
    "print(\"Total number of samples : \",len(training_data))\n",
    "print(\"Number of training samples : \",len(X_train))\n",
    "print(\"Number of testing samples : \",len(X_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Defining the learning model\n",
    " \n",
    "### MLP (Multi-Layer Perceptron)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#Defining ous model - Multi Layer Perceptron\n",
    "model=MLPClassifier(hidden_layer_sizes=(64,128,64,10),verbose=True)\n",
    "                    #(32,128,256,128,32,10))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training\n",
    "Now we will train our SVM using the training data we had split from the original dataset.(i.e - X_train , y_train )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration 1, loss = 2.61155405\n",
      "Iteration 2, loss = 1.79844934\n",
      "Iteration 3, loss = 1.49033139\n",
      "Iteration 4, loss = 1.25387473\n",
      "Iteration 5, loss = 1.07662620\n",
      "Iteration 6, loss = 0.93841767\n",
      "Iteration 7, loss = 0.81920368\n",
      "Iteration 8, loss = 0.68008650\n",
      "Iteration 9, loss = 0.54154401\n",
      "Iteration 10, loss = 0.45690169\n",
      "Iteration 11, loss = 0.38729756\n",
      "Iteration 12, loss = 0.32826706\n",
      "Iteration 13, loss = 0.27581759\n",
      "Iteration 14, loss = 0.22854166\n",
      "Iteration 15, loss = 0.19303373\n",
      "Iteration 16, loss = 0.15567661\n",
      "Iteration 17, loss = 0.12607195\n",
      "Iteration 18, loss = 0.10563596\n",
      "Iteration 19, loss = 0.09001996\n",
      "Iteration 20, loss = 0.07744777\n",
      "Iteration 21, loss = 0.06674822\n",
      "Iteration 22, loss = 0.05782977\n",
      "Iteration 23, loss = 0.05143641\n",
      "Iteration 24, loss = 0.04400343\n",
      "Iteration 25, loss = 0.03999499\n",
      "Iteration 26, loss = 0.03642126\n",
      "Iteration 27, loss = 0.03167276\n",
      "Iteration 28, loss = 0.02879737\n",
      "Iteration 29, loss = 0.02635282\n",
      "Iteration 30, loss = 0.02390033\n",
      "Iteration 31, loss = 0.02229631\n",
      "Iteration 32, loss = 0.02102651\n",
      "Iteration 33, loss = 0.01938716\n",
      "Iteration 34, loss = 0.01814796\n",
      "Iteration 35, loss = 0.01714418\n",
      "Iteration 36, loss = 0.01638923\n",
      "Iteration 37, loss = 0.01533341\n",
      "Iteration 38, loss = 0.01456711\n",
      "Iteration 39, loss = 0.01407993\n",
      "Iteration 40, loss = 0.01336342\n",
      "Iteration 41, loss = 0.01271407\n",
      "Iteration 42, loss = 0.01216384\n",
      "Iteration 43, loss = 0.01118367\n",
      "Iteration 44, loss = 0.01044571\n",
      "Iteration 45, loss = 0.00965254\n",
      "Iteration 46, loss = 0.00878954\n",
      "Iteration 47, loss = 0.00807203\n",
      "Iteration 48, loss = 0.00750234\n",
      "Iteration 49, loss = 0.00700836\n",
      "Iteration 50, loss = 0.00654546\n",
      "Iteration 51, loss = 0.00621463\n",
      "Iteration 52, loss = 0.00575001\n",
      "Iteration 53, loss = 0.00555328\n",
      "Iteration 54, loss = 0.00521993\n",
      "Iteration 55, loss = 0.00493735\n",
      "Iteration 56, loss = 0.00477917\n",
      "Iteration 57, loss = 0.00456709\n",
      "Iteration 58, loss = 0.00436767\n",
      "Iteration 59, loss = 0.00411895\n",
      "Iteration 60, loss = 0.00401006\n",
      "Iteration 61, loss = 0.00376473\n",
      "Iteration 62, loss = 0.00361854\n",
      "Iteration 63, loss = 0.00345760\n",
      "Iteration 64, loss = 0.00333526\n",
      "Iteration 65, loss = 0.00316766\n",
      "Iteration 66, loss = 0.00305211\n",
      "Iteration 67, loss = 0.00291090\n",
      "Iteration 68, loss = 0.00281582\n",
      "Iteration 69, loss = 0.00269651\n",
      "Iteration 70, loss = 0.00256056\n",
      "Iteration 71, loss = 0.00247325\n",
      "Iteration 72, loss = 0.00238173\n",
      "Iteration 73, loss = 0.00230280\n",
      "Training loss did not improve more than tol=0.000100 for two consecutive epochs. Stopping.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,\n",
       "       beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
       "       hidden_layer_sizes=(64, 128, 64, 10), learning_rate='constant',\n",
       "       learning_rate_init=0.001, max_iter=200, momentum=0.9,\n",
       "       nesterovs_momentum=True, power_t=0.5, random_state=None,\n",
       "       shuffle=True, solver='adam', tol=0.0001, validation_fraction=0.1,\n",
       "       verbose=True, warm_start=False)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fitting the training data to the model\n",
    "model.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Making prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Making predictions with out model\n",
    "prediction=model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([7, 5, 6, 0, 9, 0, 9, 2, 5, 4, 8, 0, 2, 4, 2, 7, 6, 9, 7, 7, 9, 8,\n",
       "       4, 0, 2, 2, 0, 0, 6, 0, 8, 2, 5, 1, 0, 5, 8, 1, 6, 3, 8, 5, 7, 5,\n",
       "       1, 9, 3, 9, 0, 9, 9, 7, 8, 7, 8, 3, 4, 4, 1, 4, 5, 7, 7, 9, 0, 0,\n",
       "       4, 0, 6, 7, 3, 3, 0, 7, 7, 2, 1, 8, 8, 1, 9, 2, 2, 8, 1, 1, 3, 5,\n",
       "       4, 7, 0, 1, 8, 5, 6, 6, 6, 9, 3, 2, 0, 0, 9, 1, 2, 4, 9, 0, 7, 0,\n",
       "       8, 1, 4, 6, 9, 0, 8, 4, 0, 9, 2, 8, 1, 0, 3, 4, 5, 4, 0, 8, 2, 2,\n",
       "       9, 9, 9, 5, 7, 2, 4, 1, 5, 7, 2, 3, 7, 7, 7, 1, 2, 1, 9, 3, 3, 3,\n",
       "       6, 4, 1, 4, 8, 9, 8, 6, 8, 0, 9, 0, 6, 2, 8, 8, 4, 2, 4, 2, 0, 4,\n",
       "       6, 9, 6, 9, 6, 0, 8, 8, 5, 1, 5, 1, 6, 5, 1, 5, 6, 9, 7, 7, 5, 1,\n",
       "       7, 3, 3, 3, 1, 5, 4, 5, 2, 1, 4, 9, 9, 3, 4, 3, 0, 7, 5, 2, 5, 2,\n",
       "       0, 5, 1, 7, 9, 1, 7, 2, 3, 0, 4, 6, 6, 7, 2, 5, 6, 7, 1, 0, 1, 1,\n",
       "       6, 7, 9, 7, 8, 8, 6, 2, 0, 7, 9, 6, 2, 4, 3, 4, 3, 6, 1, 3, 4, 4,\n",
       "       3, 1, 6, 3, 2, 8, 4, 3, 7, 8, 6, 1, 3, 2, 4, 5, 1, 7, 9, 0, 1, 2,\n",
       "       9, 6, 7, 6, 9, 1, 4, 9, 6, 9, 4, 4, 6, 4, 7, 9, 6, 7, 6, 5, 0, 2,\n",
       "       6, 6, 2, 4, 1, 2, 0, 8, 8, 3, 0, 8, 7, 2, 4, 6, 2, 1, 7, 3, 6, 3,\n",
       "       2, 9, 5, 4, 0, 2, 0, 3, 4, 1, 2, 6, 3, 8, 5, 6, 9, 7, 5, 0, 6, 8,\n",
       "       7, 7, 7, 9, 3, 0, 5, 4, 8, 3, 2, 1, 0, 1, 7, 1, 5, 8, 7, 3, 5, 1,\n",
       "       3, 3, 8, 2, 8, 9, 4, 5, 7, 6, 0, 2, 0, 0, 2, 9, 7, 3, 4, 4, 2, 2,\n",
       "       0, 1, 5, 6, 9, 1, 5, 4, 3, 8, 7, 3, 4, 9, 4, 3, 9, 3, 7, 4, 9, 4,\n",
       "       4, 3, 4, 8, 5, 5, 8, 7, 4, 9, 8, 2, 8, 4, 5, 5, 3, 3, 7, 8, 9, 4,\n",
       "       9, 5, 9, 0, 9, 9, 3, 9, 3, 7])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Whats all this? How to make sense of it?\n",
    "\n",
    "All these numbers stored in the prediction array!What are they? <br>\n",
    "These are the classes of the images in the X_test as <b>predicted by our MLP</b>.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([7, 5, 6, 0, 9, 0, 9, 2, 5, 4, 8, 0, 2, 4, 2, 7, 6, 9, 7, 7, 9, 8,\n",
       "       4, 0, 2, 2, 0, 0, 6, 0, 8, 2, 5, 1, 0, 5, 8, 1, 6, 3, 8, 5, 7, 5,\n",
       "       1, 9, 3, 9, 0, 9, 9, 7, 8, 7, 8, 3, 4, 4, 1, 4, 5, 7, 7, 9, 0, 0,\n",
       "       4, 0, 6, 7, 3, 3, 0, 7, 7, 2, 1, 8, 8, 1, 9, 2, 2, 8, 1, 1, 3, 5,\n",
       "       4, 7, 0, 1, 8, 5, 6, 6, 6, 9, 3, 2, 0, 0, 9, 1, 2, 4, 9, 0, 7, 0,\n",
       "       8, 1, 4, 6, 9, 0, 8, 4, 0, 9, 2, 8, 1, 0, 3, 4, 5, 4, 0, 8, 2, 2,\n",
       "       9, 9, 9, 5, 7, 2, 4, 1, 5, 7, 2, 3, 4, 7, 7, 1, 2, 1, 9, 3, 3, 3,\n",
       "       6, 4, 1, 4, 8, 9, 8, 6, 8, 0, 9, 0, 6, 2, 8, 8, 4, 2, 4, 2, 0, 4,\n",
       "       6, 9, 6, 9, 6, 0, 8, 8, 5, 1, 5, 1, 6, 5, 1, 5, 6, 9, 7, 7, 5, 1,\n",
       "       7, 3, 3, 3, 1, 5, 4, 5, 2, 1, 4, 9, 9, 3, 4, 3, 0, 7, 5, 2, 5, 3,\n",
       "       0, 5, 1, 7, 9, 1, 7, 2, 3, 0, 4, 6, 6, 7, 2, 5, 6, 7, 1, 0, 1, 1,\n",
       "       6, 7, 9, 7, 8, 8, 6, 2, 0, 7, 9, 6, 1, 4, 3, 4, 3, 6, 6, 3, 4, 4,\n",
       "       3, 1, 6, 3, 2, 8, 4, 3, 7, 8, 6, 1, 3, 2, 4, 5, 1, 7, 9, 0, 1, 2,\n",
       "       9, 6, 7, 6, 9, 1, 4, 9, 6, 9, 4, 4, 6, 4, 7, 9, 6, 7, 6, 5, 0, 2,\n",
       "       6, 6, 2, 4, 1, 2, 0, 5, 8, 3, 0, 8, 7, 2, 4, 6, 2, 1, 7, 3, 6, 3,\n",
       "       2, 9, 5, 4, 0, 2, 0, 3, 4, 1, 2, 6, 3, 8, 5, 6, 9, 7, 5, 0, 6, 8,\n",
       "       7, 9, 7, 9, 3, 0, 5, 4, 8, 3, 2, 1, 0, 1, 7, 8, 5, 8, 7, 3, 5, 1,\n",
       "       3, 3, 8, 2, 8, 9, 4, 5, 7, 6, 0, 2, 0, 0, 2, 9, 7, 3, 4, 4, 2, 2,\n",
       "       0, 8, 5, 6, 9, 1, 5, 4, 3, 8, 7, 3, 4, 3, 4, 3, 9, 2, 7, 4, 9, 4,\n",
       "       4, 3, 4, 3, 5, 5, 8, 7, 4, 9, 8, 2, 8, 4, 5, 5, 3, 3, 7, 8, 9, 4,\n",
       "       9, 5, 3, 0, 9, 7, 3, 5, 3, 7])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluating our model prediction\n",
    "\n",
    "But how do I actully verify the predictions of the model?<br>\n",
    "Remeber the y_test array?<br>It contained all the true output labels of the X_test images.<br>\n",
    "We can compare the prediction and the actual labels(y_test) to see the performace. \n",
    "\n",
    "Use the following piece of code to verify your predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Misclassified at ctr= 144 Predicted to be :  7 True Label :  4\n",
      "Misclassified at ctr= 219 Predicted to be :  2 True Label :  3\n",
      "Misclassified at ctr= 254 Predicted to be :  2 True Label :  1\n",
      "Misclassified at ctr= 260 Predicted to be :  1 True Label :  6\n",
      "Misclassified at ctr= 315 Predicted to be :  8 True Label :  5\n",
      "Misclassified at ctr= 353 Predicted to be :  7 True Label :  9\n",
      "Misclassified at ctr= 367 Predicted to be :  1 True Label :  8\n",
      "Misclassified at ctr= 397 Predicted to be :  1 True Label :  8\n",
      "Misclassified at ctr= 409 Predicted to be :  9 True Label :  3\n",
      "Misclassified at ctr= 413 Predicted to be :  3 True Label :  2\n",
      "Misclassified at ctr= 421 Predicted to be :  8 True Label :  3\n",
      "Misclassified at ctr= 442 Predicted to be :  9 True Label :  3\n",
      "Misclassified at ctr= 445 Predicted to be :  9 True Label :  7\n",
      "Misclassified at ctr= 447 Predicted to be :  9 True Label :  5\n",
      "Final Score : 436 / 450\n",
      "Number of miscalssifications : 14\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Evaluating model performance\n",
    "ctr=0\n",
    "score=0\n",
    "misclassified=[]\n",
    "for i,j in zip(y_test,prediction):\n",
    "    if i==j:\n",
    "        score+=1\n",
    "    else:\n",
    "        misclassified.append((X_test[ctr],y_test[ctr],prediction[ctr]))#Logging misclassified data\n",
    "        #misclassified.append((X_test[ctr],output_label[ctr],prediction[ctr])) \n",
    "        print (\"Misclassified at ctr=\",ctr,\"Predicted to be : \",prediction[ctr],\"True Label : \",y_test[ctr])\n",
    "    ctr+=1\n",
    "\n",
    "print (\"Final Score :\",score,\"/\",len(y_test))\n",
    "print (\"Number of miscalssifications :\",len(misclassified))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### What went wrong?\n",
    "\n",
    "Lets take a look at where the prediction went wrong "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPgAAAD8CAYAAABaQGkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAACv9JREFUeJzt3W+slnUdx/HPxwN4BEk3xXIcFjxIypqJYzRHuQnT4Z+pD2rBhlvWdnqQDpfLaetJD2r2IGcP0uZQc5N0hVLmTHOJmqtI/mnCAYek4zAVnTOQGXj024NzsaHDznW4r3/3d+/Xxjx/7p3f91bfXNe5z3WunyNCAHI6oe0BANSHwIHECBxIjMCBxAgcSIzAgcQIHEiMwIHECBxIbEodX3SaT4xBzajjS7fKU2r51/WJZn3+3cbW2vOf0xpbq0mDrx1qdL0YG2tknf/qoA7HIU/0uFr+jx3UDH3FS+v40q0aOP2MRtcb/v3fGlvrhkdWNrZWk+bfsrvR9T54Y18j62yIv5R6HKfoQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRWKnDby2zvtL3L9k11DwWgGhMGbntA0i8lXSLpbEkrbJ9d92AAelfmCL5I0q6I2B0RhyU9IOnKescCUIUygc+WtOeo90eLjwHouMp+2cT2sKRhSRrU9Kq+LIAelDmC75U056j3h4qPfURE3BkRCyNi4VSdWNV8AHpQJvDnJH3O9jzb0yQtl/RwvWMBqMKEp+gRMWb7WkmPSxqQdHdEbKt9MgA9K/U9eEQ8KunRmmcBUDGuZAMSI3AgMQIHEiNwIDECBxIjcCAxAgcSI3AgsWb34ulzr94+q9H1tr+X85f2Xv7mrxpb69JbLm5srS7iCA4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJFZmZ5O7be+z/WITAwGoTpkj+K8lLat5DgA1mDDwiHhG0tsNzAKgYnwPDiTG1kVAYpUdwdm6COgeTtGBxMr8mOx+SX+XNN/2qO3v1D8WgCqU2ZtsRRODAKgep+hAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJNb3WxcNfHF+Y2ttO39NY2tJ0te+993G1hpceaCxtS7cdmVja01749XG1uoijuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRW5qaLc2yvt73d9jbbq5oYDEDvylyLPibphojYbHumpE22n4iI7TXPBqBHZfYmey0iNhdvH5A0Iml23YMB6N2kfpvM9lxJCyRtOMbn2LoI6JjSL7LZPlnSg5Kuj4j9H/88WxcB3VMqcNtTNR73moh4qN6RAFSlzKvolnSXpJGIuLX+kQBUpcwRfLGkqyUtsb21+HNpzXMBqECZvcmeleQGZgFQMa5kAxIjcCAxAgcSI3AgMQIHEiNwIDECBxIjcCCxvt+b7N/fOK2xtV56/2Bja0nS/rkDja31k3P+0NhaV814t7G1fvpCc3vXSdLT55zU6HoT4QgOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRW5qaLg7b/afv5YuuiHzcxGIDelblU9ZCkJRHxbnH75Gdt/yki/lHzbAB6VOamiyHpyMXDU4s/UedQAKpRduODAdtbJe2T9EREHHPrItsbbW98X4eqnhPAcSgVeER8EBHnShqStMj2l47xGLYuAjpmUq+iR8Q7ktZLWlbPOACqVOZV9Fm2Ty3ePknSRZJ21D0YgN6VeRX9TEn32h7Q+F8Iv42IR+odC0AVyryK/oLG9wQH0Ge4kg1IjMCBxAgcSIzAgcQIHEiMwIHECBxIjMCBxPp+66Kh9e81ttZZwzMaW0uSnv/B7Y2t1eS2TPMeW9XYWl/4+YHG1hq3s+H1/j+O4EBiBA4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYqUDL+6NvsU292MD+sRkjuCrJI3UNQiA6pXd2WRI0mWSVtc7DoAqlT2C3ybpRkkf1jgLgIqV2fjgckn7ImLTBI9jbzKgY8ocwRdLusL2K5IekLTE9n0ffxB7kwHdM2HgEXFzRAxFxFxJyyU9GREra58MQM/4OTiQ2KTu6BIRT0l6qpZJAFSOIziQGIEDiRE4kBiBA4kROJAYgQOJETiQGIEDifX91kUnPL2lsbUuWnFNY2tJ0tiP3m5srVf2zGpsrbO+vbGxtT5obKVu4ggOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRW6kq24o6qBzR+YdBYRCyscygA1ZjMpaoXRsRbtU0CoHKcogOJlQ08JP3Z9ibbw3UOBKA6ZU/RvxoRe22fIekJ2zsi4pmjH1CEPyxJg5pe8ZgAjkepI3hE7C3+uU/SOkmLjvEYti4COqbM5oMzbM888rakiyW9WPdgAHpX5hT905LW2T7y+N9ExGO1TgWgEhMGHhG7JX25gVkAVIwfkwGJETiQGIEDiRE4kBiBA4kROJAYgQOJETiQWN9vXdSkJrdJkqS5P5vZ2Fpjf/xMY2uhORzBgcQIHEiMwIHECBxIjMCBxAgcSIzAgcQIHEiMwIHESgVu+1Tba23vsD1i+/y6BwPQu7KXqv5C0mMR8XXb0yRufA70gwkDt32KpAskfUuSIuKwpMP1jgWgCmVO0edJelPSPba32F5d3B8dQMeVCXyKpPMk3RERCyQdlHTTxx9ke9j2Rtsb39ehiscEcDzKBD4qaTQiNhTvr9V48B/B1kVA90wYeES8LmmP7fnFh5ZK2l7rVAAqUfZV9OskrSleQd8t6Zr6RgJQlVKBR8RWSQtrngVAxbiSDUiMwIHECBxIjMCBxAgcSIzAgcQIHEiMwIHECBxIjL3JIEmavm7DxA9C3+EIDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kNmHgtufb3nrUn/22r29iOAC9mfBS1YjYKelcSbI9IGmvpHU1zwWgApM9RV8q6eWIeLWOYQBUa7K/bLJc0v3H+oTtYUnDkjTI5qNAJ5Q+ghebHlwh6XfH+jxbFwHdM5lT9EskbY6IN+oaBkC1JhP4Cn3C6TmAbioVeLEf+EWSHqp3HABVKrs32UFJp9U8C4CKcSUbkBiBA4kROJAYgQOJETiQGIEDiRE4kBiBA4k5Iqr/ovabkib7K6WnS3qr8mG6Ietz43m157MRMWuiB9US+PGwvTEiFrY9Rx2yPjeeV/dxig4kRuBAYl0K/M62B6hR1ufG8+q4znwPDqB6XTqCA6hYJwK3vcz2Ttu7bN/U9jxVsD3H9nrb221vs72q7ZmqZHvA9hbbj7Q9S5Vsn2p7re0dtkdsn9/2TL1o/RS9uNf6Sxq/Y8yopOckrYiI7a0O1iPbZ0o6MyI2254paZOkq/r9eR1h+/uSFkr6VERc3vY8VbF9r6S/RsTq4kaj0yPinbbnOl5dOIIvkrQrInZHxGFJD0i6suWZehYRr0XE5uLtA5JGJM1ud6pq2B6SdJmk1W3PUiXbp0i6QNJdkhQRh/s5bqkbgc+WtOeo90eVJIQjbM+VtEDShnYnqcxtkm6U9GHbg1RsnqQ3Jd1TfPuxurgfYd/qQuCp2T5Z0oOSro+I/W3P0yvbl0vaFxGb2p6lBlMknSfpjohYIOmgpL5+TagLge+VNOeo94eKj/U921M1HveaiMhyR9rFkq6w/YrGv51aYvu+dkeqzKik0Yg4cqa1VuPB960uBP6cpM/Znle8qLFc0sMtz9Qz29b493IjEXFr2/NUJSJujoihiJir8f9WT0bEypbHqkREvC5pj+35xYeWSurrF0UnuzdZ5SJizPa1kh6XNCDp7ojY1vJYVVgs6WpJ/7K9tfjYDyPi0RZnwsSuk7SmONjslnRNy/P0pPUfkwGoTxdO0QHUhMCBxAgcSIzAgcQIHEiMwIHECBxIjMCBxP4H3d+ZZYT/HMcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 4\n",
      "Predicted class : 7\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 2\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPgAAAD8CAYAAABaQGkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAACplJREFUeJzt3d2PXWUZhvH77tAytiBNEJS01TaxNiESKGlKsEq0DaYIAQ44aBOIEpM5gkAkIhgP9B8geGBIsIAkVFALREIQJFK+olbaaVXaaU1tkE4DFkKwH2JL4fFgVpNCamZN97s+5vH6JRNm79mZ9eyBi7X2njXrdUQIQE4zuh4AQHMIHEiMwIHECBxIjMCBxAgcSIzAgcQIHEiMwIHETmvim87y6TGsOU186059OLfd5zT8mfda29bpM461tq2Dr7f4czzU3s+wTf/RYR2NI57scY0EPqw5usSrmvjWnfr3ykta3d7nv7ejtW0tnr2/tW1tvPlLrW1rxgtbW9tWmzbF72o9jkN0IDECBxIjcCAxAgcSI3AgMQIHEiNwIDECBxKrFbjt1bZ32d5t+46mhwJQxqSB2x6S9BNJV0g6X9Ja2+c3PRiAwdXZgy+XtDsi9kTEUUmPSLqm2bEAlFAn8HmS9p5we7y6D0DPFftjE9sjkkYkaVizS31bAAOoswffJ2nBCbfnV/d9RETcGxHLImLZTJ1eaj4AA6gT+CuSFtteZHuWpDWSnmh2LAAlTHqIHhHHbN8k6RlJQ5Luj4jtjU8GYGC1XoNHxFOSnmp4FgCFcSYbkBiBA4kROJAYgQOJETiQGIEDiRE4kBiBA4k1srJJVm9fONTq9m47u71VOa6dc6i1bf30m19pbVtfeKG1TfUSe3AgMQIHEiNwIDECBxIjcCAxAgcSI3AgMQIHEiNwILE6K5vcb3u/7VfbGAhAOXX24D+TtLrhOQA0YNLAI+JFSe+0MAuAwngNDiTG0kVAYsX24CxdBPQPh+hAYnV+TfawpD9IWmJ73Pa3mx8LQAl11iZb28YgAMrjEB1IjMCBxAgcSIzAgcQIHEiMwIHECBxIjMCBxFi6aAo++8Pft7q9X69e2tq2rp3zUmvbmvEv/rNrC3twIDECBxIjcCAxAgcSI3AgMQIHEiNwIDECBxIjcCAxAgcSq3PRxQW2N9reYXu77VvaGAzA4OqcFHxM0m0RMWr7TElbbD8bETsang3AgOqsTfZGRIxWnx+UNCZpXtODARjclP6sx/ZCSUslbTrJ11i6COiZ2m+y2T5D0qOSbo2IAx//OksXAf1TK3DbMzUR9/qIeKzZkQCUUudddEu6T9JYRNzV/EgASqmzB18h6QZJK21vqz6+0fBcAAqoszbZy5LcwiwACuNMNiAxAgcSI3AgMQIHEiNwIDECBxIjcCAxAgcSY5GoHhv9xQXtbey77a1Ndsbr7Ffawk8aSIzAgcQIHEiMwIHECBxIjMCBxAgcSIzAgcQIHEiszkUXh23/yfafq6WLftTGYAAGV+dU1SOSVkbEoeryyS/b/k1E/LHh2QAMqM5FF0PSoermzOojmhwKQBl1Fz4Ysr1N0n5Jz0bESZcusr3Z9ub3daT0nABOQa3AI+KDiLhI0nxJy21/8SSPYekioGem9C56RLwraaOk1c2MA6CkOu+in2N7bvX5JyRdLmln04MBGFydd9HPk/Sg7SFN/A/hlxHxZLNjASihzrvof9HEmuAAphnOZAMSI3AgMQIHEiNwIDECBxIjcCAxAgcSI3AgMZYu6rEDFxztegRMc+zBgcQIHEiMwIHECBxIjMCBxAgcSIzAgcQIHEiMwIHEagdeXRt9q22uxwZME1PZg98iaaypQQCUV3dlk/mSrpS0rtlxAJRUdw9+t6TbJX3Y4CwACquz8MFVkvZHxJZJHsfaZEDP1NmDr5B0te3XJD0iaaXthz7+INYmA/pn0sAj4s6ImB8RCyWtkfRcRFzf+GQABsbvwYHEpnRFl4h4XtLzjUwCoDj24EBiBA4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kxtJFPfbV83d1PUIjzh19r+sR/m+wBwcSI3AgMQIHEiNwIDECBxIjcCAxAgcSI3AgMQIHEqt1Jlt1RdWDkj6QdCwiljU5FIAypnKq6tci4u3GJgFQHIfoQGJ1Aw9Jv7W9xfZIkwMBKKfuIfqXI2Kf7XMlPWt7Z0S8eOIDqvBHJGlYswuPCeBU1NqDR8S+6p/7JT0uaflJHsPSRUDP1Fl8cI7tM49/Lunrkl5tejAAg6tziP5pSY/bPv74n0fE041OBaCISQOPiD2SLmxhFgCF8WsyIDECBxIjcCAxAgcSI3AgMQIHEiNwIDECBxJj6aIeWzx7f9cjYJpjDw4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJFYrcNtzbW+wvdP2mO1Lmx4MwODqnqr6Y0lPR8R1tmdJXPgcmA4mDdz2WZIuk/QtSYqIo5KONjsWgBLqHKIvkvSWpAdsb7W9rro+OoCeqxP4aZIulnRPRCyVdFjSHR9/kO0R25ttb35fRwqPCeBU1Al8XNJ4RGyqbm/QRPAfwdJFQP9MGnhEvClpr+0l1V2rJO1odCoARdR9F/1mSeurd9D3SLqxuZEAlFIr8IjYJmlZw7MAKIwz2YDECBxIjMCBxAgcSIzAgcQIHEiMwIHECBxIjMCBxFibrMeuO2u0xa219xfAx37wTmvbmvVCa5vqJfbgQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiRG4EBikwZue4ntbSd8HLB9axvDARjMpKeqRsQuSRdJku0hSfskPd7wXAAKmOoh+ipJf4+IfzQxDICypvrHJmskPXyyL9gekTQiScMsPgr0Qu09eLXowdWSfnWyr7N0EdA/UzlEv0LSaET8s6lhAJQ1lcDX6n8cngPop1qBV+uBXy7psWbHAVBS3bXJDks6u+FZABTGmWxAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJOaIKP9N7bckTfVPSj8l6e3iw/RD1ufG8+rO5yLinMke1Ejgp8L25ohY1vUcTcj63Hhe/cchOpAYgQOJ9Snwe7seoEFZnxvPq+d68xocQHl92oMDKKwXgdtebXuX7d227+h6nhJsL7C90fYO29tt39L1TCXZHrK91faTXc9Sku25tjfY3ml7zPalXc80iM4P0atrrf9NE1eMGZf0iqS1EbGj08EGZPs8SedFxKjtMyVtkXTtdH9ex9n+jqRlkj4ZEVd1PU8pth+U9FJErKsuNDo7It7teq5T1Yc9+HJJuyNiT0QclfSIpGs6nmlgEfFGRIxWnx+UNCZpXrdTlWF7vqQrJa3repaSbJ8l6TJJ90lSRBydznFL/Qh8nqS9J9weV5IQjrO9UNJSSZu6naSYuyXdLunDrgcpbJGktyQ9UL38WFddj3Da6kPgqdk+Q9Kjkm6NiANdzzMo21dJ2h8RW7qepQGnSbpY0j0RsVTSYUnT+j2hPgS+T9KCE27Pr+6b9mzP1ETc6yMiyxVpV0i62vZrmng5tdL2Q92OVMy4pPGIOH6ktUETwU9bfQj8FUmLbS+q3tRYI+mJjmcamG1r4rXcWETc1fU8pUTEnRExPyIWauLf1XMRcX3HYxUREW9K2mt7SXXXKknT+k3Rqa5NVlxEHLN9k6RnJA1Juj8itnc8VgkrJN0g6a+2t1X3fT8inupwJkzuZknrq53NHkk3djzPQDr/NRmA5vThEB1AQwgcSIzAgcQIHEiMwIHECBxIjMCBxAgcSOy/Z1R+oLELMC8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 1\n",
      "Predicted class : 2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 6\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 5\n",
      "Predicted class : 8\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 9\n",
      "Predicted class : 7\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPgAAAD8CAYAAABaQGkdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAACs1JREFUeJzt3duLXeUZx/HfzzExRmNi1VrNhCZQCVipRkLEpghNUGIVleJFAhEqhbQXitYWq95I/wGxF8Ui8QSmShsVRKxWPAua5tjWnGya2maiNh6qiSnNGH16MSsQJWXWzn7XWnsevx8IzmEz77Mx36w1e9as1xEhADkd0/UAAJpD4EBiBA4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kdmwTX3Syj4spOqGJL92p0TPbfU6nzNjX2lonD/2ntbX+/tbpra019P7+1tZq03+1X6NxwOM9rpHAp+gEXeDFTXzpTv3zR99udb1rvv9sa2tdPX1Da2stv/0nra118v2vtrZWm9ZEvb8bnKIDiRE4kBiBA4kROJAYgQOJETiQGIEDiRE4kFitwG0vsb3d9g7btzQ9FIAyxg3c9pCkX0m6VNLZkpbZPrvpwQD0r84RfIGkHRGxMyJGJT0s6cpmxwJQQp3AZ0raddj7I9XHAAy4Yr9sYnuFpBWSNEVTS31ZAH2ocwTfLWnWYe8PVx/7nIi4OyLmR8T8STqu1HwA+lAn8LWSzrI9x/ZkSUslPd7sWABKGPcUPSIO2r5O0tOShiTdGxGbG58MQN9qfQ8eEU9KerLhWQAUxpVsQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGIEDiTWyM4mWY1O/6zV9W47dXuLq7W3LdM3f/x6a2u9dX9rSw0kjuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYgQOJEbgQGJ1dja51/Ye2+1dfgSgiDpH8PslLWl4DgANGDfwiHhJ0gctzAKgML4HBxJj6yIgsWJHcLYuAgYPp+hAYnV+TPaQpFclzbU9YvuHzY8FoIQ6e5Mta2MQAOVxig4kRuBAYgQOJEbgQGIEDiRG4EBiBA4kRuBAYmxd1IPJH+X99/CNT/a3ttbIz7/R2lrHaGNraw2ivH9jARA4kBmBA4kROJAYgQOJETiQGIEDiRE4kBiBA4kROJBYnZsuzrL9vO0ttjfbvqGNwQD0r8616Acl/TQiNtieJmm97WciYkvDswHoU529yd6OiA3V2/skbZU0s+nBAPSvp98msz1b0jxJa47wObYuAgZM7RfZbJ8o6RFJN0bE3i9+nq2LgMFTK3DbkzQW96qIeLTZkQCUUudVdEu6R9LWiLij+ZEAlFLnCL5Q0jWSFtneVP35XsNzASigzt5kr0hyC7MAKIwr2YDECBxIjMCBxAgcSIzAgcQIHEiMwIHECBxIjL3JejD5o64naM7y23/W2lonv/hqa2t92XEEBxIjcCAxAgcSI3AgMQIHEiNwIDECBxIjcCAxAgcSq3PTxSm2/2j7T9XWRb9oYzAA/atzqeoBSYsi4uPq9smv2P59RLzW8GwA+lTnposh6ePq3UnVn2hyKABl1N34YMj2Jkl7JD0TEUfcusj2OtvrPtGB0nMCOAq1Ao+ITyPiPEnDkhbYPucIj2HrImDA9PQqekR8KOl5SUuaGQdASXVeRT/N9ozq7eMlXSxpW9ODAehfnVfRz5D0gO0hjf2D8NuIeKLZsQCUUOdV9D9rbE9wABMMV7IBiRE4kBiBA4kROJAYgQOJETiQGIEDiRE4kBhbF/XguH/n/S3ZU9d+0Npan7a2EjiCA4kROJAYgQOJETiQGIEDiRE4kBiBA4kROJAYgQOJ1Q68ujf6Rtvcjw2YIHo5gt8gaWtTgwAor+7OJsOSLpO0stlxAJRU9wh+p6SbJX3W4CwACquz8cHlkvZExPpxHsfeZMCAqXMEXyjpCttvSnpY0iLbD37xQexNBgyecQOPiFsjYjgiZktaKum5iFje+GQA+sbPwYHEerqjS0S8IOmFRiYBUBxHcCAxAgcSI3AgMQIHEiNwIDECBxIjcCAxAgcSY+uiHrx/bt6ti3Zfckpra31tc2tLfelxBAcSI3AgMQIHEiNwIDECBxIjcCAxAgcSI3AgMQIHEqt1JVt1R9V9kj6VdDAi5jc5FIAyerlU9bsR8V5jkwAojlN0ILG6gYekP9heb3tFkwMBKKfuKfp3ImK37a9Kesb2toh46fAHVOGvkKQpmlp4TABHo9YRPCJ2V//dI+kxSQuO8Bi2LgIGTJ3NB0+wPe3Q25IukfR604MB6F+dU/TTJT1m+9DjfxMRTzU6FYAixg08InZKOreFWQAUxo/JgMQIHEiMwIHECBxIjMCBxAgcSIzAgcQIHEjMEeW34znJX4kLvLj41+3ama9N63qExpw1dU9ra734reNbWyurNfGs9sYHHu9xHMGBxAgcSIzAgcQIHEiMwIHECBxIjMCBxAgcSIzAgcRqBW57hu3VtrfZ3mr7wqYHA9C/uvdF/6WkpyLiatuTJW58DkwE4wZue7qkiyT9QJIiYlTSaLNjASihzin6HEnvSrrP9kbbK6v7owMYcHUCP1bS+ZLuioh5kvZLuuWLD7K9wvY62+s+0YHCYwI4GnUCH5E0EhFrqvdXayz4z2HrImDwjBt4RLwjaZftudWHFkva0uhUAIqo+yr69ZJWVa+g75R0bXMjASilVuARsUnS/IZnAVAYV7IBiRE4kBiBA4kROJAYgQOJETiQGIEDiRE4kBiBA4nVvVQVkjb/+pxW17vqpudaW2vVX9u7UHFYm1tb68uOIziQGIEDiRE4kBiBA4kROJAYgQOJETiQGIEDiRE4kNi4gduea3vTYX/22r6xjeEA9GfcS1UjYruk8yTJ9pCk3ZIea3guAAX0eoq+WNLfIuIfTQwDoKxef9lkqaSHjvQJ2yskrZCkKWw+CgyE2kfwatODKyT97kifZ+siYPD0cop+qaQNEfGvpoYBUFYvgS/T/zk9BzCYagVe7Qd+saRHmx0HQEl19ybbL+mUhmcBUBhXsgGJETiQGIEDiRE4kBiBA4kROJAYgQOJETiQmCOi/Be135XU66+UnirpveLDDIasz43n1Z2vR8Rp4z2okcCPhu11EdHeBlktyvrceF6Dj1N0IDECBxIbpMDv7nqABmV9bjyvATcw34MDKG+QjuAAChuIwG0vsb3d9g7bt3Q9Twm2Z9l+3vYW25tt39D1TCXZHrK90fYTXc9Sku0Ztlfb3mZ7q+0Lu56pH52folf3Wn9DY3eMGZG0VtKyiNjS6WB9sn2GpDMiYoPtaZLWS7pqoj+vQ2zfJGm+pJMi4vKu5ynF9gOSXo6IldWNRqdGxIddz3W0BuEIvkDSjojYGRGjkh6WdGXHM/UtIt6OiA3V2/skbZU0s9upyrA9LOkySSu7nqUk29MlXSTpHkmKiNGJHLc0GIHPlLTrsPdHlCSEQ2zPljRP0ppuJynmTkk3S/qs60EKmyPpXUn3Vd9+rKzuRzhhDULgqdk+UdIjkm6MiL1dz9Mv25dL2hMR67uepQHHSjpf0l0RMU/SfkkT+jWhQQh8t6RZh70/XH1swrM9SWNxr4qILHekXSjpCttvauzbqUW2H+x2pGJGJI1ExKEzrdUaC37CGoTA10o6y/ac6kWNpZIe73imvtm2xr6X2xoRd3Q9TykRcWtEDEfEbI39v3ouIpZ3PFYREfGOpF2251YfWixpQr8o2uveZMVFxEHb10l6WtKQpHsjYnPHY5WwUNI1kv5ie1P1sdsi4skOZ8L4rpe0qjrY7JR0bcfz9KXzH5MBaM4gnKIDaAiBA4kROJAYgQOJETiQGIEDiRE4kBiBA4n9D/FJkaDaYHDkAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 8\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 8\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 9\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 2\n",
      "Predicted class : 3\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 8\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 9\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 7\n",
      "Predicted class : 9\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 5\n",
      "Predicted class : 9\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Looking at our misclassification\n",
    "for data in misclassified:\n",
    "    plt.imshow(data[0].reshape(8,8))\n",
    "    plt.show()\n",
    "    print (\"Expected class :\",data[1])\n",
    "    print (\"Predicted class :\",data[2])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## So is this the best the MLP can perform?\n",
    "\n",
    "We can expect better results from our network by quite a few ways but lets take a look at what we are'nt doing right->\n",
    "\n",
    "#### Challenges\n",
    "* It is usually noticed that Neural Networks perform best when the volume of our data is large.(our dataset is only about 1700 samples, which isn't as big in the ML world).\n",
    "\n",
    "\n",
    "* As we convert the the 2-D image into a 1-D vector for training purpose our dimensional corelation is lost.This problem was overcome by using a new type of Neural Networks called Convolutional Neural Nets(CNN). \n",
    "    (Ref :  __[CNN Paper - Yann LeCun ](http://yann.lecun.com/exdb/publis/pdf/lecun-01a.pdf)__ )\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ML != Neural Nets \n",
    "There are loads of other algorithms depending on the learning problem one is trying to solve.\n",
    "But yes,NN have proven to out-perform most other algorithm in cases which involes large datasets(which in today todays world is the most often case)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### What other choice do we have?\n",
    "   We can look into out ml-module(sci-kit learn) documentation for help!\n",
    "  __[Model Map](http://scikit-learn.org/stable/tutorial/machine_learning_map/index.html)__\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lets try a SVM?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import svm\n",
    "\n",
    "#Defining ous model - Support Vector Machine\n",
    "svm_model=svm.SVC(gamma=0.001)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training the SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibSVM]"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma=0.001, kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=True)"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fitting the training data to the model\n",
    "svm_model.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Making Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Making predictions with our model\n",
    "svm_prediction=svm_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Misclassified at ctr= 49\n",
      "Misclassified at ctr= 151\n",
      "Misclassified at ctr= 315\n",
      "Misclassified at ctr= 360\n",
      "Misclassified at ctr= 367\n",
      "Misclassified at ctr= 397\n",
      "Misclassified at ctr= 442\n",
      "Misclassified at ctr= 445\n",
      "Misclassified at ctr= 447\n",
      "Score : 441 / 450\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Evaluating model performance\n",
    "ctr=0\n",
    "score=0\n",
    "misclassified=[]\n",
    "for i,j in zip(y_test,svm_prediction):\n",
    "    if i==j:\n",
    "        score+=1\n",
    "    else:\n",
    "      misclassified.append((X_test[ctr],y_test[ctr],svm_prediction[ctr])) #Logging misclassified data\n",
    "      print (\"Misclassified at ctr=\",ctr)\n",
    "    ctr+=1\n",
    "\n",
    "print (\"Score :\",score,\"/\",len(y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 9\n",
      "Predicted class : 7\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 2\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 5\n",
      "Predicted class : 6\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 8\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 8\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 8\n",
      "Predicted class : 1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 3\n",
      "Predicted class : 7\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 7\n",
      "Predicted class : 9\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Expected class : 5\n",
      "Predicted class : 9\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#Looking at our misclassification\n",
    "for data in misclassified:\n",
    "    plt.imshow(data[0].reshape(8,8))\n",
    "    \n",
    "    plt.show()\n",
    "    print (\"Expected class :\",data[1])\n",
    "    print (\"Predicted class :\",data[2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}