hand_written_digit_identification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Load Mnist dataset.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting python-mnist\n",
      "  Downloading https://files.pythonhosted.org/packages/cd/e5/05e37dcf2ed8966a618cd348721e68a2bfad302438344acc16ed94e349fd/python-mnist-0.5.tar.gz\n",
      "Building wheels for collected packages: python-mnist\n",
      "  Running setup.py bdist_wheel for python-mnist ... \u001b[?25ldone\n",
      "\u001b[?25h  Stored in directory: /home/nbuser/.cache/pip/wheels/e3/67/aa/bea4c3dc84f590c0dabb8c00317b6b7f8e6bd43b095d7679d5\n",
      "Successfully built python-mnist\n",
      "Installing collected packages: python-mnist\n",
      "Successfully installed python-mnist-0.5\n",
      "\u001b[33mYou are using pip version 9.0.3, however version 10.0.1 is available.\n",
      "You should consider upgrading via the 'pip install --upgrade pip' command.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install python-mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from mnist import MNIST\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load Train and test images and labels in different objects\n",
    "mndata = MNIST('../../data')\n",
    "train_images, train_labels = mndata.load_training()\n",
    "test_images, test_labels = mndata.load_testing()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are 55k examples in train, and 10k in eval. You may wish to limit the size to experiment faster. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Limit the number of training data by using max_examples.\n",
    "max_examples = 20000\n",
    "train_data = np.asarray(train_images[:max_examples], dtype=np.float)\n",
    "train_labels = np.asarray(train_labels[:max_examples], dtype=np.int32)\n",
    "\n",
    "test_data = np.asarray(test_images, dtype=np.float)\n",
    "test_labels = np.asarray(test_labels, dtype=np.int32)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Display using 0,1 matrix.\n",
    "def display_matrix(obj,d):\n",
    "    img = obj[d].reshape(28,28)\n",
    "    for i in range(28):\n",
    "        print('')\n",
    "        for j in range(28):\n",
    "            if img[i][j] > 0:\n",
    "                print(1,end=\"\")\n",
    "            else:\n",
    "                print(0,end=\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000000000000000000000000000\n",
      "0000001111110000000000000000\n",
      "0000001111111111111111000000\n",
      "0000001111111111111111000000\n",
      "0000000000011111111111000000\n",
      "0000000000000000001111000000\n",
      "0000000000000000011110000000\n",
      "0000000000000000011110000000\n",
      "0000000000000000111100000000\n",
      "0000000000000000111100000000\n",
      "0000000000000001111000000000\n",
      "0000000000000001110000000000\n",
      "0000000000000011110000000000\n",
      "0000000000000111100000000000\n",
      "0000000000001111100000000000\n",
      "0000000000001111000000000000\n",
      "0000000000011111000000000000\n",
      "0000000000011110000000000000\n",
      "0000000000111110000000000000\n",
      "0000000000111110000000000000\n",
      "0000000000111100000000000000\n",
      "0000000000000000000000000000"
     ]
    }
   ],
   "source": [
    "display_matrix(test_data,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Display using pixel \n",
    "def display(data,label,i):\n",
    "    img = data[i]\n",
    "    plt.title('Example %d. Label: %d' % (i, label[i]))\n",
    "    plt.imshow(img.reshape((28,28)), cmap=plt.cm.gray_r)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f66ded4fa90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(test_data,test_labels,209)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a dictionary of the digits (0-9), add each instance of the digit vector in corresponsing \n",
    "# value of the digit.\n",
    "# data = { 0:[[<784 pixels of first instance of 0>],[....]....[]],\n",
    "#           1:[[<784 pixels of first instance of 1>],[....]....[]],\n",
    "#              :\n",
    "#              :\n",
    "#               9:[[<784 pixels of first instance of 9>],[....]....[]],\n",
    "#       }\n",
    "\n",
    "data = {}\n",
    "for i in range(len(train_labels)):\n",
    "    key = train_labels[i]\n",
    "    if key in data:\n",
    "        data[key].append(train_data[i])\n",
    "    else:\n",
    "        data[key]=[train_data[i]]\n",
    "# Create a dictionary of mean and variance of each of the pixels of the digits ( 0 - 9)   \n",
    "mean_var_prior = {}\n",
    "for key in data:\n",
    "    mean_var_prior[key] = [np.mean(data[key],axis=0)]\n",
    "    mean_var_prior[key].append(np.std(data[key],axis=0))\n",
    "    mean_var_prior[key].append(len(data[key])/len(train_labels))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a function to derive the pdf using mean and variance. \n",
    "# Return 1.0 in case of 0.0 mean or standard deviation.\n",
    "def pdf(x,mean,std):\n",
    "    ret = 0.0\n",
    "    if std != 0.0 and mean != 0.0:\n",
    "        ret =  (np.exp(- ((x - mean)**2 / (2 * std**2)))) / (np.sqrt(2 * np.pi) * std)\n",
    "    if ret == 0.0:\n",
    "        return 1.0\n",
    "    else:\n",
    "        return ret\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test data is : 7, Predicted data is : 9\n",
      "Test data is : 2, Predicted data is : 2\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 0, Predicted data is : 0\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 9\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 9\n",
      "Test data is : 0, Predicted data is : 0\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 9, Predicted data is : 9\n",
      "Test data is : 0, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 9\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 3, Predicted data is : 2\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 0, Predicted data is : 0\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 0, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 4, Predicted data is : 0\n",
      "Test data is : 7, Predicted data is : 5\n",
      "Test data is : 2, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 2, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 7, Predicted data is : 9\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 2, Predicted data is : 5\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 5\n",
      "Test data is : 2, Predicted data is : 6\n",
      "Test data is : 4, Predicted data is : 9\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 3, Predicted data is : 3\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 0, Predicted data is : 5\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 9, Predicted data is : 9\n",
      "Test data is : 5, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 8, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 5\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 4, Predicted data is : 4\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 0, Predicted data is : 0\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 0, Predicted data is : 0\n",
      "Test data is : 2, Predicted data is : 5\n",
      "Test data is : 9, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 2, Predicted data is : 2\n",
      "Test data is : 9, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 7\n",
      "Test data is : 7, Predicted data is : 9\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 2, Predicted data is : 2\n",
      "Test data is : 7, Predicted data is : 9\n",
      "Test data is : 8, Predicted data is : 5\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 9\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 9, Predicted data is : 5\n",
      "Test data is : 3, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 1\n",
      "Test data is : 4, Predicted data is : 5\n",
      "Test data is : 1, Predicted data is : 5\n",
      "Test data is : 7, Predicted data is : 8\n",
      "Test data is : 6, Predicted data is : 6\n",
      "Test data is : 9, Predicted data is : 9\n",
      "0.53\n"
     ]
    }
   ],
   "source": [
    "count = 0\n",
    "limit = 100\n",
    "for index in range(limit):\n",
    "    pro = {}\n",
    "    num = []\n",
    "    # Interate over the digits.\n",
    "    for i in range(10):\n",
    "        p  = 0\n",
    "        # Interate over the pixels, pdf values are converted to respective log for ease of comparision\n",
    "        # the pdf values are between 0 to 1, which is difficult to compare, so it is \n",
    "        # scaled into log values.\n",
    "        for j in range(784):\n",
    "            p  = p + np.log(pdf(test_data[index][j],mean_var_prior[i][0][j],mean_var_prior[i][1][j]))\n",
    "        pro[i] = p * mean_var_prior[i][2]\n",
    "    num = sorted(pro.items(), key=lambda value: value[1])\n",
    "    print(\"Test data is :\",test_labels[index],end=\", \")\n",
    "    print(\"Predicted data is :\",num[-1][0])\n",
    "    if test_labels[index] == num[-1][0]:\n",
    "        count = count + 1\n",
    "        \n",
    "print(count/limit)\n"
   ]
  }
 ],
 "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}