{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Information Gain\n",
    "\n",
    "Nbviewer [https://nbviewer.jupyter.org/github/shaundsouza/deep-learning/blob/master/ig.ipynb](https://nbviewer.jupyter.org/github/shaundsouza/ai-ecosystems-enabling/blob/master/ig.ipynb)\n",
	"\n",
    "Execute on Binder [https://mybinder.org/v2/gh/shaundsouza/ai-ecosystems-enabling/master?filepath=ig.ipynb](https://mybinder.org/v2/gh/shaundsouza/ai-ecosystems-enabling/master?filepath=ig.ipynb)\n",
    "\n",
    "Code [https://nbviewer.jupyter.org/format/script/github/shaundsouza/ai-ecosystems-enabling/blob/master/ig.ipynb](https://nbviewer.jupyter.org/format/script/github/shaundsouza/deep-learning/blob/master/ig.ipynb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "def B(q):\n",
    "    if q == 0 or q == 1:\n",
    "        return 0\n",
    "    else:\n",
    "        return -(q * np.log2([q])[0] + (1 - q) * np.log([1 - q])[0])\n",
    "\n",
    "def H(q):\n",
    "    x = pd.unique(q)\n",
    "    h = 0\n",
    "    for i in x:\n",
    "        p = sum(q == i) / len(q)\n",
    "        # print(i, p)\n",
    "        h = h + p * np.log2([p])[0]\n",
    "    return -h\n",
    "\n",
    "def remainder(q, t):\n",
    "    x = pd.unique(q)\n",
    "    h = 0\n",
    "    for i in x:\n",
    "        # print(i)\n",
    "        pknk = sum(q == i)\n",
    "        pk = sum(t[q == i] == \"Yes\")\n",
    "        p = pknk / len(q) * B(pk / pknk)\n",
    "        # print(pknk, pk, p)\n",
    "        h = h + p\n",
    "    return h\n",
    "\n",
    "def remainderH(q, t):\n",
    "    x = pd.unique(q)\n",
    "    h = 0\n",
    "    for i in x:\n",
    "        # print(i)\n",
    "        pknk = sum(q == i)\n",
    "        # pk = sum(t[q == i] == \"Yes\")\n",
    "        p = pknk / len(q) * H(t[q == i])\n",
    "        # print(i, pknk, p)\n",
    "        h = h + p\n",
    "    return h\n",
    "\n",
    "def gain(q, t):\n",
    "    b = B(sum(t == \"Yes\") / len(t))\n",
    "    return b - remainder(q, t)\n",
    "\n",
    "def gainH(q, t):\n",
    "    return H(t) - remainderH(q, t)\n",
    "\n",
    "dataset = pd.read_csv(\"house/train.csv\")\n",
    "size = dataset.shape\n",
    "\n",
    "dataset = dataset.fillna(0)\n",
    "names = dataset.columns.values\n",
    "\n",
    "for i in range(size[1]):\n",
    "    for j in range(size[1]):\n",
    "        print(names[i], names[j], gainH(dataset.iloc[:, i], dataset.iloc[:, j]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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
}