Assignments/A3/A3_Farsi_Hand_written_digits.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load the data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The original data is taken from \n",
    "\n",
    "http://farsiocr.ir/\n",
    "\n",
    "website and transformed to the current format. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.load('X.npz')['arr_0']\n",
    "Y = np.load('Y.npz')['arr_0']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rand_ind = np.random.randint(0,100000)\n",
    "fig = plt.figure(figsize=(6,4))\n",
    "ax = fig.add_subplot(111,title = 'This should be a {}'.format(Y[rand_ind]))\n",
    "ax.imshow(X[rand_ind])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.reshape(-1, 32*32)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Assignment\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Train a some basic classifiers to get a sense of the bias and variance.\n",
    "- Calculate the confusion matrix. (You can use the built-in scikit-learn functions)\n",
    "- Calculate the recall, precision, AUC and f1-score for each class. (You can use the built-in scikit-learn functions)\n",
    "- Do you need more data?\n",
    "- Make a comparison table like the one in \n",
    "http://yann.lecun.com/exdb/mnist/  \n",
    "\n",
    "5 Classifier would suffice. You can use one type of estimator but with different settings, e.g. kernels. Also calculate the accuracy for the training and test set with cross-validation and add training time and prediction time for the test set to your table.\n",
    "\n",
    "- Apply PCA and TSNE for n_component = 2, 100\n",
    "- Find the (roughly) optimal value of n_components for PCA. \n",
    "- Train your classifiers in the comparison table, this time with the reduced data. \n",
    "\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": "3.6.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}