01-perceptrons.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Perceptrons"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Import libraries\n",
    "%matplotlib inline\n",
    "from sklearn import datasets\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# import some data to play with\n",
    "iris = datasets.load_iris()\n",
    "X = iris.data[:100,:2]  # we only take the first two features.\n",
    "Y = iris.target[:100]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Line separating the features. Change this to get lines separating the features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Line parameters\n",
    "m = 1.1 # Slope\n",
    "c = -2.75  # Intercept\n",
    "\n",
    "# Creating line\n",
    "x1 = np.linspace(np.min(X[:,0]), np.max(X[:,0]), 100)\n",
    "y1 = m*x1 + c"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Let's plot this!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7fe54d593f10>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfcAAAF/CAYAAABUloCSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd4FNX3x/H3paT3RoDQe5VOEIXQBQREUSkKKKIiWBFQ\nFAHLTxELFopdVMQCCuLXAiKRIr1HCEVqqIGQBAiQdn5/DCIqCSHs7iSb83qefdxNdvZ+RsSzM3Pn\nHiMiKKWUUsp9FLM7gFJKKaUcS4u7Ukop5Wa0uCullFJuRou7Ukop5Wa0uCullFJuRou7Ukop5WZK\nOHsAY8weIAXIBjJEpNm/ft8amAvsOv+jb0TkeWfnUkoppdyV04s7VlGPEZETubxnsYh0d0EWpZRS\nyu254rS8ycM4xgU5lFJKqSLBFcVdgJ+NMauNMYNzeE+0MWa9MeZ/xpjaLsiklFJKuS1XnJZvKSKH\njDHhwAJjzFYRWXrR79cCFUQkzRjTGZgDVHdBLqWUUsotGVeuLW+MGQucFJHXcnnPbqCxiCT96+e6\nCL5SSqkiR0Su+NK1U0/LG2N8jDF+55/7Ah2BuH+9p9RFz5thfeH4R2H/i4gU2cfYsWNtz6D7r/uv\n+677r/vv2kd+Ofu0fCng2/NH3SWAGSIy3xhzHyAi8i7QyxgzBMgAzgC3OzmTUkop5dacWtxFZDfQ\n4BI/f+ei55OByc7MoZRSShUlukJdIRETE2N3BFvp/sfYHcE2RXnfQfe/qO9/frl0Qt3VMMZIYcmq\nlFJKOYIxBiloE+qUUkop5Xpa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8Vd\nKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3\no8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVS\nSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa3JVSSik3o8VdKaWUcjNa\n3JVSSik3o8VdKaWUcjNOL+7GmD3GmI3GmPXGmFU5vOdNY8wOY8wGY0wDZ2dSSiml3FkJF4yRDcSI\nyIlL/dIY0xmoIiLVjDHNgWlAtAtyKaWUUm7JFaflzWXG6QF8AiAiK4FAY0wpF+RSSiml3JIrirsA\nPxtjVhtjBl/i92WB/Re9PnD+Z0oppZTKB1eclm8pIoeMMeHAAmPMVhFZ6oJxlVJKqSLJ6cVdRA6d\n/2eiMeZboBlwcXE/AJS76HXU+Z/9x7hx4y48j4mJISYmxsFplVJKKfvExsYSGxt71Z9jROTq0+T0\n4cb4AMVE5JQxxheYD4wXkfkXvacLMFREuhpjooFJIvKfCXXGGHFmVqWUUqqgMcYgIuZKt3P2kXsp\n4FtjjJwfa4aIzDfG3AeIiLwrIj8YY7oYY3YCp4G7nJxJKaWUcmtOPXJ3JD1yV0qpouWLLyA6GipW\ntDuJffJ75K4r1CmllCpQTp6EAQNg7Fg4dcruNIWTFnellFIFxpo10KgRlCwJ69ZB3bp2JyqcXHEr\nnFJKKZWr7Gx45RXrMXky3Hqr3YkKNy3uSimlbHXoEPTvD2fPwurVUKGC3YkKPz0tr5RSyjbffw8N\nG8L118OiRVrYHUWP3JVSSrnc2bMwYgTMmwezZsF119mdyL3okbtSSimX2rIFmjWDo0dhwwYt7M6g\nxV0ppZRLiMC0adC6NTzyiHUfe1CQ3anck56WV0op5XTHj8M998DevbB0KdSoYXci96ZH7koppZxq\n0SJo0ACqVIHly7Wwu4IeuSullHKKjAxrlbmPP4YPP4QbbrA7UdGhxV0ppZTD7doFfftCSAisXw+l\nStmdqGjR0/JKKaUcasYMaN4ceve27mPXwu56euSulFLKIVJTYdgwWLUK5s+3FqdR9tAjd6WUUldt\n1Sqr4YunJ6xdq4XdbnrkrpRSKt+ys+Hll+G112DKFOjVy+5ECrS4K6WUyqeDB+HOO61Z8WvWQPny\ndidSf9HT8koppa7Yd99Zp+FjYqz72LWwFyx65K6UUirPzpyBxx+HH36A2bOhZUu7E6lL0SN3pZRS\neRIXZzV8OX7cunddC3vBpcVdKaVUrkSsyXJt2sBjj8HMmdrwpaDT0/JKKaVydOyY1fBl/35Ytgyq\nV7c7kcoLPXJXSil1Sb/+at2vXq2a1fBFC3vhoUfuSiml/iEjA555Bj75BD76CDp2tDuRulJa3JVS\nSl3w55/Qpw+Eh1uT5iIi7E6k8kNPyyullALg008hOhr69bMavmhhL7z0yF0ppYq41FR44AFrTfhf\nfoFrrrE7kbpaeuSulFJF2MqV1qQ5X1+ruGthdw965K6UUkVQVhZMmACTJsHUqXDLLXYnUo6kxV0p\npYqYAweshi9ZWdbRerlydidSjqan5ZVSqgiZO9dq+NKmjXUfuxZ296RH7kopVQScOQPDh8OPP8Kc\nOdCihd2JlDPpkbtSSrm5zZuhaVM4cQI2bNDCXhS4pLgbY4oZY9YZY767xO8GGGOOnv/9OmPM3a7I\npJRS7k4E3n4b2raFESPg888hMNDuVMoVXHVa/mFgCxCQw++/EJGHXJRFKaXc3rFjcPfdcOgQ/P67\ntT68KjqcfuRujIkCugDv5/Y2Z+dQSqmi4pdfoEEDqFnT6uSmhb3occWR++vACCC3k0E3G2OuB7YD\nj4lIggtyKaWUW0lPhzFj4LPPYPp0aN/e7kTKLk49cjfGdAWOiMgGrKPzSx2hfwdUFJEGwC/AdGdm\nUkopd7RzJ7RsCX/8YU2a08JetDn7yL0l0N0Y0wXwBvyNMZ+ISP+/3iAiJy56//vAyzl92Lhx4y48\nj4mJISYmxtF5lVKqUBGxWrM+/rjVpnXYMDB6obPQio2NJTY29qo/x4jI1afJy0DGtAaGi0j3f/08\nUkQOn3/eExghItdeYntxVVallCoMUlJgyBDYuBFmzoT69e1OpBzNGIOIXPHXNVvuczfGjDfG3Hj+\n5UPGmDhjzHpgGDDQjkxKKVWYrFhhNXwJDITVq7Wwq39y2ZH71dIjd6WUstaDf+klePNNmDYNeva0\nO5Fypvweuevys0opVUgkJFgNX0Sshi9RUXYnUgWVLj+rVBG3evVq7r5nMHcNuodly5bZHUfl4Ntv\noXFjaxb8woVa2FXutLgrVYStWLGCTjd0Jj2wNFkh5ejW4yYWLVpkdyx1kbQ0a9Lc8OFWR7ennoLi\nxe1OpQo6veauVBHWu09fvCvUpuPtAwFYPO9r9q+J5fvv5tobTAGwaRP06WOtNjdliq4LXxQVqtny\nSqmCIT0jAy8f3wuvvXx8SU9PtzGRAuua+ltvQbt28MQTMGOGFnZ1ZXRCnVJF2N0DB3D34Hvx8Q+g\nePESfDHpeV5/Jcd1pJQLJCZaDV+OHIHly6FqVbsTqcJIT8srVcTNmjWLN956GxFhyH330q9fP7sj\nFVkLFsDAgdC/P4wfDx4edidSdsvvaXkt7kopZbP0dHj6aavf+vTp1ul4pUDvc1dKqUJpxw5r0lyZ\nMlbDl7AwuxMpd6AT6pRSygYi8PHHcO21cNdd1m1uWtiVo+iRu1JKuVhKCtx/P2zeDL/+CvXq2Z1I\nuRs9cldKKRdavty6bz042Gr4ooVdOYMeuSullAtkZcGLL8Lbb8M770CPHnYnUu5Mi7tSSjnZ/v1w\nxx3WsrFr10LZsnYnUu5OT8srpZQTffMNNGkCN9xg3ceuhV25gh65K6WUE6SlwaOPwi+/wHffQfPm\ndidSRYkeuSullINt3Gi1Zz19Gtav18KuXE+Lu1JKOYgIvPmm1XP9qafgs88gIMDuVKoo0uKuVAG1\ne/duYtq0pVRkaa5vHcOOHTvsjqRycfQo3Hij1cFtxQprAp1SdtHirlQBdO7cOdp36EhUg2sZ89Fc\nKjdrQ/sOHUlLS7M7mrqE+fOhYUO45hpYuhSqVLE7kSrqtLgrVQDFx8eTZYrRtf/9hEaW4Ya+g/Dw\n8SMuLs7uaOoi6ekwYgQMGgSffgr/939QsqTdqZTS2fJKFUgBAQGknkjibNppvHx8OXfmDMlJxwjQ\nC7gFxvbtVsOXqChr0pyuC68KEm35qlQBdc/ge1myYjX1WrbljxWxNKlfl0+mf4wxV9z9UTnQXw1f\nRo60eq4PGQL6R6KcRfu5K+VmRISZM2cSFxdH7dq16du3L8WK6ZU0OyUnw333wR9/wBdfQN26didS\n7k6Lu1JKOdGyZdCvnzUjfuJE8Pa2O5EqCvJb3PWau1JK5SIz05ooN2UKvPsudO9udyKlLk+Luyry\nUlNTee+99zh+/DgdOnSgTZs2dkdSBcS+fdbRuoeHNnxRhYtewFNF2qlTp4i+tiVzflnCjhPp9O53\nBx9+9JHdsVQBMGuW1fCla1frPnYt7Kow0Wvuqkh77733eH/mLB559QMA9sTH8cZjd3P40EGbkym7\nnD4NjzwCv/4KM2dCs2Z2J1JFWX6vueuRuyrSTp48SUipMhdeh0aW5dSpUzYmUnbasME6Wk9Pt+5d\n18KuCist7qpI69SpEyvnz2P9koUc2b+H6ROeolv3bnbHUi6WnQ2TJkHHjjBmDEyfrg1fVOGmp+VV\nkbdgwQIee3wkSUnWhLrJb72Jr6+v3bGUixw5AgMHwokT8PnnULmy3YmU+pve566UUlfo55/hrrvg\n7rth7FhdF14VPAX6PndjTDFgDZAgIt3/9TsP4BOgMXAMuF1E9rkil1KqaDp3DkaPhq++slq06t2P\nyt246pr7w8CWHH43CEgSkWrAJOBlF2VSSuVTRkYG27dv5+jRo3ZHuWLbtkGLFrBrlzWBTgu7ckdO\nL+7GmCigC/B+Dm/pAUw//3wW0M7ZmZRS+bdr1y5q1alLm/YdqVKtGiNGjqIwXDITgQ8+gOuug3vv\nhW++gdBQu1Mp5RyuOC3/OjACCMzh92WB/QAikmWMSTbGhIhIkguyKaWuUP+BdxHd9Va69r+fUykn\neGHwrbS6/jq6dSu4dxmcOGEV9G3bIDYW6tSxO5FSzuXUI3djTFfgiIhsAMz5x2U3c2YmpdTVidu8\nmeu73QqAX2AwDa5vz8aNG21OlbOlS6FBAyhdGlat0sKuigZnH7m3BLobY7oA3oC/MeYTEel/0XsS\ngHLAQWNMcSAgp6P2cePGXXgeExNDTEyMs3IrpXJQuUoV1i9ZSOvut5F+9gxb1yyjd8dRdsf6j8xM\neP55mDYN3nsPCvCJBaUuiI2NJTY29qo/x2W3whljWgPDLzFb/gGgrog8YIzpDdwkIr0vsb3eCqdU\nAbBp0yY6drqB8DLlSDx8kHZt2/Dp9I8LVK/5vXuthi/e3taCNGXKXH4bpQqiAn0r3L8ZY8YDq0Xk\ne+AD4FNjzA7gOPCfwq6UKjjq169P/NYtbNy4kaCgIOrXr48xBedq2tdfw9Ch8Pjj1qMAfedQymV0\nERullFs4fRoeeggWL7ZWmmva1O5ESl09bRyjlCqy1q2DRo0gK8t6roVdFXVa3JVygbNnz9KqdQwh\n4RFElo3ik08+sTuSW8jOhtdeg06dYNw4+Phj8Pe3O5VS9rPlmrtSRU2LltdxOhMenDCNQ3v+ZPB9\n91G2bFnatdM1m/Lr8GGr4UtqqnWLW6VKdidSquDQa+5KuYCHlxevz11CSERpAN5/bhR+WaeYO3eu\nzckKpx9/tJq93HOP1fClhB6mKDdVqGbLK1XUFCtWjDOnT194ffpUCiGB3jYmKpzOnYMnnoDZs+GL\nL6B1a7sTKVUwaXFXygU6dOjAS0PvoOc9D5Gwazsbly3iozVr7I5VqMTHQ58+1un3DRsgJMTuREoV\nXDqhTikXmDd3Ln169eS3WR+TuG0jy5YsoVatWnbHKhRE4P334frr4f77raN2LexK5U6vuSulCqy/\nGr5s3w4zZ0Lt2nYnUsq19D53pZRbWbzYavhSpgysXKmFXakrodfcldu5//77WbVqFc2aNWPatGl2\nxynwjh07xnfffYeIcOONN1KqVClb82RmwrPPWs1e3n8funa1NY5ShdJlT8sbYzyBW4CKXPRlQESe\ndWqy/+bQ0/LqsspXrMSptDPUb9Gajb/HEuDny97du+yOVWDt37+f6BbXUqluQ4oVK8b29atYtnQJ\nVapUsSXPnj1WwxdfX6vhS+nStsRQqsDI72n5vBT3n4AUYC2Q9dfPReTVKx3samhxV5fz9ddf03/g\nXbz5v+X4BQZzMvkED3WNZuaMz+jZs6fd8Qqkwffex3Hx5NYHRgLw3YdvI8f38cXnn7s8y5dfwoMP\nwsiR8Nhj2vBFKXDufe5RInJDPjIp5VJxcXEEhobjFxgMgH9QMIGh4cTFxWlxz8HhI0eo3PLvv97l\nqtdm9RzX3qJ36pRV1Jctgx9+gCZNXDq8Um4pL9+NfzfG1HN6EqWuUr9+/TiReITff55LVmYmy378\nluRjR+nbt6/d0QqsDu3a8vOM90g5nsjJ5BP8+Ok02rdr67Lx1661Gr4YYzV80cKulGPkeFreGLMZ\nEKyj+2rALuAcYAARkfquCnk+j56WV5f1/PPP88KLL3H2TBpe3j488/RTPPnkk3bHKrCys7MZMXIU\nU6dMQRDuvnsQb74xieLFizt5XKvhy8svw5tvQu/eTh1OqULL4dfcjTEVcttQRPZe6WBXQ4u7uhLp\n6el4eHjYHaPQEBFEhGIuuNB9+DAMGGCdjp8xAypWdPqQShVaDr/PXUT2ni/gz//1/OKfXU1YpZxN\nC/uVMca4pLD/8AM0bAjR0fDbb1rYlXKWvEyoq3PxC2NMcaCxc+IodfVEhNOnT+Pr64sxV/yFN8/S\n09MRETw9PfO8TVpaGp6enk4/7V3QnD1rNXz59ltrVnyrVnYnUsq95fhV3RjzpDHmJFDfGJN6/nES\nOApon0pVIP3++++UjSpHWFg4ZaPK8fvvvzt8jMzMTAbdMxg/f3/8AwIYeNfdZGZm5rrN4cOHadHy\nOoJDQvD3D2DylCkOz1VQbd1qHaknJFgNX7SwK+V8uZ2Wf1FE/IGJIhJw/uEvIqEiojOUVIGTkpJC\nj5t60m/k83y0fAd9RzxHj549SU1Ndeg4L0+cyJq4eKYt3Mg7v25i47Y/efGll3Ld5s7+AwmvVo8P\nl23n/76cz3Mv/B9LlixxaK6CRgTefdcq5kOHwtdfQ3Cw3amUKhpyO3JvZIxpBHz91/OLHy7MqFSe\nbN++neDwUjRq1R6Axq07EBQawfbt2x06zuIlS2l32wC8ff3w8vGl3W0DWbxkaa7bLF/+O93uGkqx\nYsUoFVWBpu1vdMpZhYIiKQl69YKpU2HJEhg82LrdTSnlGrnNoHn1/GMysBJ4F3jv/PPJzo+m1JWJ\njIzk6MEEUo4nApB87CiJhw5Q2sFrmEaVLcuuzesvvN4Vt46yZcvmnq10aXae3yY7K4s9WzZSpkwZ\nh+YqKH77zWr4Uq4crFgBNWvanUipoicvy89+A4wVkc3nX9cFxolILxfkuziH3gqnLuuFF/6Pt6ZM\noWbDZmxdt5KHhw1j9GjHXkU6ePAg17a8jrCoihhjOLpvF78vW5prgV+0aBG39LqVus2v48j+PVQo\nW5ofvp9HiRLu07spMxPGj7eavXzwAXTpYncipQo/Z64t/4eI/HvG/H9+5mxa3FVerVq1ivj4eGrW\nrEmzZs2cMkZKSgrz589HROjYsSNBQUGX3Wb37t0sW7aMkJAQOnXq5FYz5vfsgb59wd/favgSGWl3\nIqXcgzOL+0zgNPDZ+R/1A/xEpM8Vp7wKWtyVKphmzoSHH4ZRo+DRR7Xhi1KO5Mzi7gUMAf66gWUx\nMFVEzl5xyqugxV2pguXkSavhy/LlVoFvpNNslXI4h69Q9xcROSsir4tIz/OP111d2FXeiQiTp0yh\nR8+bufe++9m/f7/dkfJNRJg6bRo9et7MPYPvZe9el654rHKxZo1VzIsX/7v5i1Kq4MjtVrivzv9z\nszFm078frouorsSTo5/ijSnvUKF5e45lexDd4loSExPtjpUvY8eO47W3plC+eTtOGB+iW1zLkSNH\n7I5VpGVnw8SJ1mS555+3Js75+dmdSin1b7k1jiktIodyaiCjjWMKHhHBPyCQCbMWEhJh3f41ZfRQ\n+t98I4MHD7Y53ZULDglh/Kc/EF4mCoBpYx6md9f2DBkyxOZkRdOhQ9C/P6SlacMXpVzFGY1jDp1/\n2h7wuETzGFUASXY2xUuUvPC6RMmSZGVl2Zgo/7KysilR8u9bxQrzvhR2339vNXxp2VIbvihVGORl\nQt144HqgIrAWa0LdEhHZ4PR0/8yhR+558OBDDxO7fBVdBzzA/p3bWPDF+2xYt65QLpjy6GPDmf/b\nUrrdNYwDu3fw02fvsH7tWqKiouyOVmScPQsjR8LcufDZZ3D99XYnUqpocdps+YsG8AYGA48DZUXE\npTfpanHPm6ysLCa8/DLzF/xCeFg4Lzz/LNWrV7c7Vr5kZWUx8ZVX+Onn+YSFhfH8s+OpqcuducyW\nLdCnD1Svbq0Rr+vCK+V6zrwV7mmgJeAHrAeWYh25H8p1QwfT4q6Ua/zV8OXpp+Gll+Duu3VdeKXs\nkt/inpe1L28GMoH/Ab8By0XkXB5DeWKdxvc4P9YsERn/r/cMACYCCed/9LaIfJi3+EopR0pKgnvu\ngd27YelSqFHD7kRKqfzIy33ujbAm1a0COgCbjTG5t8D6e9tzQBsRaQg0ADobYy61HugXItLo/EML\nu7pgw4YNjBkzhnfeeYfs7Ow8bfP9998zevRovv/++zy9X0RYtmwZ8+bN4/Dhw1cT1+Hi4+OZO3cu\nW7ZscfpYsbFWw5eKFa2GL1rYlSrERCTXB1AXa4W6L4CdwCLg2cttd4nP8QHWAE3/9fMBwFt52F5U\n0TJlyhTx8PKWmo2aS0ip0lKxSlXJyMjIdZubb+kl3r7+UrtJC/H29ZNet96W6/uzsrKk1223S7lK\nVaRpq7YSEhYmy5Ytc+Ru5Nvbb0+WkLBwada6vYSER8hrr7/ulHHS00VGjxYpXVrkxx+dMoRSKp/O\n174rqrcikqdr7t9jnVpfCqwWkYwr+fJgjCmGNcu+CjBZRJ781+8HAP8HJALbgcdEJOESnyOXy6rc\ni19AIPc/+zqNW3ckI/0cT/XtQu+buzNx4sRLvn/dunW0aNmSV2bHEla6LIkHExjRqw0rly+nQYMG\nl9zmq6++YszzL/LUe7Mo6eHJ6kU/8f27r7I9fqszd+2yjh49StXq1Xnusx+IKFue44cP8nTfG4jb\nvMmhdwvs2gX9+kFQEHz8MZQq5bCPVko5gNOuuYvIjfmLdGH7bKChMSYAmGOMqS0iF59j/A74XEQy\njDH3AtOBdpf6rHHjxl14HhMTQ0xMzNVEUwXcmbTT1GnaEoCSHp7UahLN9u3bc3z/xo0bCS1VhrDS\nVuvV8DJRhIRHsmnTphyL+969e6l2TVNKengCUKdpS6Y9/bCD9+TKJSQkEB5Zloiy5QEIjSxDZLkK\nJCQkOKy4f/651fBl9Gjrn9rwRSn7xcbGEhsbe9Wf47Jm0iKSaoxZBNwAbLno5ycuetv7wMs5fcbF\nxV25v6DgEH7+4iO63zWUE4mHWfXLD4x/5ukc3x8TE8PxBx5g69rl1Grcgi1rlpOUeITWrVvnuE2T\nJk14/c236XzHYILDI/nlq+k0bNzYGbtzRapWrUry8aP8sWoZdZq1JH7dSo4e3E+1atWu+rNPnoRh\nw2DlSpg/31qcRilVMPz7wHX8+PE5vzk3+TmXn9cHEAYEnn/ujXV6v8u/3hN50fOewO85fJZDrl+o\nwmPJkiXiFxAoXr5+UrxECenardtlt5kwYYJ4eHqJj1+AlPT0lAkTJlx2m5cmTBBvH18JDg2TWnXq\nyp49exwR/6otXLhQQsLCJCQ8QoJDQuTnn3++6s9ctUqkShWRQYNETp1yQEillFPhrGvuV8MYUw/r\nNHux848vReSF86verRaR740x/wd0BzKAJGCIiPzn3Ktecy+aMjMz2bTJus4cERGRp21OnTrF1q1b\nqVWrFn557Gpy+vRpUlJSiIyMpFgBOj+dkZHBkSNHiIiIwMPDI9+f81fDl1dfhSlToFcvB4ZUSjmN\nwxexMcbMA3KspiLS/UoHuxpa3JXKn4MHrYYv585ZDV/Kl7c7kVIqr5wxoe6Vq8ijlCoA5s2DwYNh\nyBB46iko4bJZNkopO+XWFe633B6uDKny7sSJE9x08y0EBgVTuWo1fvjhB6eMM2LECHz9AvDw9KJS\n1WocPHjQ4WPs27ePipWr4OHpha9/AE899dRlt9m6dStNm0cTEBhE0+bRxMfHX3abuXPnUrFyFYKC\nQ7jl1ttISUlxRPx/SEhIoFKVata++AUwatQoh49xsbNn4cEHrcfs2TB2bOEr7D/++CNVK5Qj0N+X\nHjd2Jikpye5IShUal724aIypZoyZZYzZYozZ9dfDFeHUlbuj/wBOFfPi5W9i6Tviee7oP4C4uDiH\njvHZZ58xeeo0Hn/zY978YQWRVWrRslXOM9Lz67pWMZSr3YC3flzJ8Nc/4LVJb/DVV1/l+P60tDQ6\ndOzENe1v4pU5S6jfrjsdOnYiLS0tx202bNjAXYPuof/oCUyYvYgT6YaBdw9yyr6UqV6HN39YweNv\nfMRbk6cwY8YMh48D8Mcf0LQpHD0KGzZYbVoLmy1bttDv9lsZUMXwdocyZP65ln6332p3LKUKjbzM\nHPoImIq1vnwb4BPgM2eGUvkjIiyY/zP9HhtLQHAodZq1pGnbzixatMih48yYMYPWPW6nZsNmBIaE\nMXDUcxzYv8+hYwAcPLCfgSOfIyA4lFqNW9CyS08+//zzHN+/ZcsWvP0DaNfrDvyDgml/a388fPxy\nPXpfuHAh0R27U6txNAHBofR5bAzzf/rJ4ftyIGE/A0Y+R2BIGDUbNef6br0cXtxFYOpUiImBxx6D\nL76wFqcpjGJjY4mO8uOaSF8CvUowsF4wC2N/Q+fdKJU3eSnu3iKyEGvy3V4RGQd0dW4slR/GGAKD\ngji8bzfL6a9QAAAgAElEQVRgFfsj+3YT7OBenaGhoST8uf3C/2gP7d1FiZL5n8mdk5IlPTi0zzpJ\nJCIc3L2TsLCwHN8fHBzM8cSjnDl9CoC0Uyc5cexorvsfHBzMkf27L+zL4b27CXJCb9OSHh4c2vv3\nvhzYtSPXfblSx4/DzTfD++/DsmVw112Fu5NbcHAwh05n/f3f2Ml0/H19MYV5p5RypcvdKwf8jvUl\n4BtgGNa96Nvyc9/d1TzQ+9zzZPr06RIWESnd+t8nja5rI02bR8vZs2cdOsaRI0ckIChY6kW3ks59\n7xFvX395bPhwh44hIjJ06FDx8QuQzv0GS51mLSUwOESOHz+e6zb3D3lAqtSsI90GDpHKNWrLA0OH\n5fr+tLQ0adi4iTRp1V5uvPM+CQ0vJTNnznTkboiIyCOPPCLevv7Sue89Urf59RIQFCyJiYkO+exf\nfxWJihIZPlzEwX/Utjl79qxEN2kkTSuESc/a4RIe6Ccfffih3bGUcjmcuLZ8U2ArEAQ8BwQCL4vI\nCmd94cghh1wuq7IsX76c3377jbCwMO644w68vLwcPsaxY8d45JFHOHbsGH379qV///4OHwPggw8+\n4OuvvyYiIoJJkyYREhKS6/tFhDlz5ly4z/2mm2667NHemTNn+Oyzzzh+/Dht27alWbNLNS68etOn\nT2fmzJmEhYUxadKkqz5yz8iwJspNnw4ffQQdOzooaAFx9uxZZsyYQWJiIq1ateLaa6+1O5JSLufw\n+9wvMUAA1jeIk1c6iCNocVfqb3/+CX37QliYVdjzuL6PUqqQyW9xz8ts+SbGmM3AJqxe7huNMfYv\nvq1UETVjBkRHW8X9+++1sCul/isvd75+CDwgIksAjDHXYc2gr+/MYEoB/O9//2P+/AWEhYUybNiw\ny04OzMzM5L333uOPLVupW6c2gwcPpnjx4i5K61ypqTB0KKxZA7/8AtdcY3eioik7O5uPPvqIDWvX\nUK1mLYYMGULJkiXtjqXUP+RltnzWX4UdQESWYt0Wp5RTTZ48mcFDhnKsmB+L1v5Bs+bRpKam5vh+\nEaF3335M/fgzkj2CmPzhJ/Tu288tbp9atQoaNQIfH1i7Vgu7nQYN7M/rY0eStuIbPnn1WXp07Ux2\ndrbdsZT6h7xMqJuE1dFtJtZa87cDZzl/r7uIrHNyxr9y6DX3IiYsIoJRU78gqnJ1AN4ccS+Det/M\nPffcc8n3x8fH06pNWyZ+uxgPTy/Sz53l8ZuuZ9ni36hevboroztMVpbV8OW116x72G+5xe5ERduB\nAweoU6Ma73Quh3fJYmRmCw//cpg5Py+kcQFoFazcjzPWlv/LX8cIY//184ZYxb7tlQ6qVF6cSTtD\nYEj4hdcBIWG5rjaXlpaGn38gHp7W3QElPTzx9Q/IdZuC7MABuPNOyMy0TsVrwxf7paWl4e1REq8S\n1v9rSxQzBHiXLLT/jSn35dSWr46kR+5FT/+Bd7Ft3yFuGfI4B3bt4NOXx7Bi+e85HoWfO3eOetc0\n4Jo2XWjWviurFnzPpt9+YtOG9Xh6ero4/dWZOxfuu8+6xj56NLjJtIFCLysri0b161KtWBIx5X1Z\nc+gMsUcNcfHb89xeWKkr4czZ8qWMMR8YY348/7q2Mcbxi28r9S/vTptKk9rVePepYayY8ynzvpub\n6+l1T09Pfv1lAWcSdjDtifs5c2AnCxfML1SF/cwZq6A/8gh88w2MGaOFvSApXrw483+NJatiE16P\nS+dQcA1+XbxUC7sqcPJyzf1HrNnxT4nINcaYEsB6EannioAX5dAjd+XW4uKgd2+oVw+mTYPAQLsT\nKaXs5rQjdyBMRL4CsgFEJBPIutKBlFKXJgJTpkCbNvD44/D551rYlVJXJy8T6k4bY0KxJs9hjIkG\nHN/wuojYvXs3ycnJ1KxZE29vb6eMcfr0abZt20ZYWBjl8zgLa9u2baxevZoWLVpQpUoVp+TKj7S0\nNLZt20ZISAgVKlSwO47DHTsGgwZBQoLV8KWQTurPk6SkJHbv3k358uUJDw+//AZuJiEhgaNHj1K9\nenU9jZ8H586dIz4+Hj8/PypXrqxNg67U5RafBxoBy7AK+jJgO1A/PwvZX82DQt44Jjs7W+4f8oAE\nh4ZJpWo1pULFSrJt2zaHj7Nx40YpXaasVK5RWwKDgmXkqCcuu80DQ4eKh6eXlIqqIB6eXjJi5EiH\n58qPuLg4KRtVTipVryWBwSHy2PDHJTs72+5YDrNwoUjZsiIjRoicO2d3Guf69ttvJcjfV6qVCZVA\nPx/59JNP7I7kUmNGPymBvt5StXSolAoLkbVr19odqUDbs2ePVK1YXiqVCpbQAF/p36+PZGVl2R3L\nFjircQzA+evsNQCD1REuwzlfNXLNIHnJWlDNmjWLUWPGMfrdr/Hx8+fnmR+ybdl8li9b6tBx6tSr\nT6tb76JVt1s5lXKCZ++6iQ/fnUaHDh0u+f5NmzbRtHlznvtkHuWq1mTXlk08O+gW4rduoVKlSg7N\ndqWuadiI5t370uam3pxKTeb5QTcz7a036Ny5s625rlZGBjzzDHzyCXz8MeTwR+M2UlJSqFiuLE+3\nCKNaqDf7Us7x9OIjbNm2gzJlytgdz+kWL15Mn57deKl1BIFeJVi8N5U5B4qzc88+u6MVWB3atKJM\n8nZuqRXM2cxsxv9+jJEvvMaAAQPsjuZyDr/mboxpaoyJhAvX2RsDLwCvGmNyb82l/mPLli3Uv7YN\nPn7+AER36k781i0OHUNE2LZ1Cy06dQfALzCYus1bsWVLzuMsX76ciDLlKVe1JgCVa9cnMDScVatW\nOTRbfmyL30qLjt0A8AsIuuy+FAY7d0LLlrB5M6xf7/6FHWDfvn0E+3hSLdS6DFU+0JNyIb78+eef\nNidzjS1btlC/lA+BXtZV0Jbl/Nm9L4HMTF3oMydbt8bTspwvAF4litEotBhxmzfbnKpwyW1C3TtA\nOoAxphXwEvAJ1un5d50fzb3UrFmTuJWLOZt2GoDVC3+geo2aDh3DGEO1GjVZ+csPAJw+mcIfq5dS\ns2bO4zRv3pyjB/dzYNcOAPZu30Ly8aNOa3t6JapVr8Gqhda+pJ1M5Y9Vue9LQffpp9CiBdxxB8yb\nV3QavpQrV46k02fZlXQWgITUcyScSKNy5co2J3ONmjVrEnf0DKnnrHnIKxJOUiGqDCVK5GXKU9FU\no0YNlidYCwOdy8xmQ1I2tevUsTlVIZPT+Xpg40XPJwPjLnq9IT/XAK7mgRtccx90z2AJCY+QqrXq\nSlS58rJ161aHj7Nu3TopFVlaqtepL8GhYfLoY8Mve5168L33iYenl5SpWEU8PL3kkUcecXiu/Ni4\ncaNEli4j1WrXk+DQMHnwoYcL5TX3lBSRfv1EatUS2bDB7jT2+PqrryTI31dqlg2TQF8f+fDDD+yO\n5FJPjBwhwX4+UqNsmISHBMuqVavsjlSg/fnnn1KpXFmpWjpUwgP9pM9tvfSau6OuuRtj4oAGIpJp\njIkH7hWRxX/9TkTqOv2bxz/zSE5ZC5MdO3aQnJxM7dq18fX1dcoYJ0+eZOvWrYSFheX56GjTpk2s\nWrWKFi1aUKcAfUM+deoUW7ZsITQ0tEDN4s+rFSugXz/r9Ptrr1mNX4qqxMRE/vzzTypWrEhkZKTd\ncVxu9+7dHD16lFq1ahEQEGB3nALvzJkz/PHHH/j5+VGjRo0iO1s+v9fccyvuTwFdgGNAeaCRiIgx\npiowXURaXk3gK+UuxV0VDVlZMGECvPGG1fDl5pvtTqSUKowcXtzPf2g0UBqYLyKnz/+sOuAnLuoG\nd1EWLe6qUEhIsBq+iFjX2cuVszuRUqqwcsoKdSKyQkS+/auwn//ZdlcXdpV3Z8+e5cGHHqZm7bpc\n37qN02a9L1u2jGuvu55adeox/PERpKenX3ab6BbX4hcYhF9gMNddf71Tctltzhxo3BjatYOFC7Ww\nq4Jt27ZtVKkQRaC3B2UjQlm8eLFTxvnyyy9pXK8O9WpW4/VXX0EP1JxPu8K5mTv7D2B7whFuGvwI\n+3bG89UbL7B61UqHzkyOj4/n2pbX0W/EeEqXr8ysKS/TrF4tpk2dkuM2bdu2ZUPcVob935uIwNuj\nhxHdtDE//vCDw3LZKS0Nhg+Hn36ylo9t0cLuRErlLjMzk4jgAKJLe9KpShBrD51i1pYT7Nyzz6Hr\nD/z0008M6HMrDzQIxquE4d3NqTzy5FgefPgRh43hzpy5trwqJESEr7/+isHjXqVizbq0urEXDVt1\n4KeffnLoOPPmzaN5p+5c26kHlWrV455nJvLlV1/mus26jZsYMHI8tRq3oHaTFtzx2BhWrF7j0Fx2\n2bQJmjaFlBTYsEELuyocVqxYQXr6Oe5rXIpKwV70qh1GKb+SfPHFFw4d5/NPp3NzVV8alvalVrgP\nA2v7M+OTjx06hvovLe5uxsvLm5Mnki68Ppmc5PA17L29vTmdfOLC69QTSXh75T6GwfwjV+qJJAr7\n3FcRePtt6xT8qFEwY4Y2fFGFR1BQEJlZwtlM64xoVrZw+lyWw2fy+/j6kpqefeF1yrksfIrybSMu\noqfl3czEV17hzclTaXvrAA78Gc+ezWtZt2a1Q//CJiUl0bBRY2o2a0Wp8pX45auPefrJUTwwZEiO\n2zz55JO8NukNut81lOzsLL6fPpWnR49mzJgxDsvlSomJcPfdcPgwzJwJVavanUipK1ezamUykw7R\ntnIAaw6c5sC5Euw/fBQPDw+HjREfH8910c1pG+WBVwnD/3an8eXsObRv395hY7gzp8yWL0i0uOfd\nN998w/wFvxAREc4jDz9MSIjjVwtOTExk0htvkJR0gq5dOnPjjTdedpsJEyYw6Y03ARjx+HAee+wx\nh+dyhYULYcAA6/71554DB/5/UCmXyszMZOCAAWxct4aKVaoy4/OZTrkHf/v27bwzbSoZ587R987+\nREdHO3wMd1Ugi7sxxhNYDHhgtZedJSLj//UeD6xlbRtj3VN/u4j8p6OCFndlt/R0GDMGPvsMpk8H\nPfBQSjlbfou7Uxc3FpFzxpg2IpJmjCkOLDPG/CgiF9+fNQhIEpFqxpjbgZeB3s7MpdSV2rED+vaF\nyEhr0lwRbEeulCpEnN65QETSzj/1PD/evw+/ewBjzz+fBbzt7EyOICLMnTuXHTt2UK9ePW644YbL\nbpOWlsbMmTNJTk6mffv2XHPNNS5I6hxJSUkMHz6cY8eO0bdvX/r06XPZbQ4cOMA333yDMYZbbrmF\n0qVLX3abtWvXEhsbS2hoKL1798bLy8sR8fNMxGrN+vjjMHYsDB0K+V0Fc/369fz6668EBwfTp08f\nh090BEhNTeWxxx7j8OHD3HbbbfTv3/+y2xw+fJhZs2YhIvTs2ZOoqCiH5wL48ccfiYuLo3r16nTv\n3t0py4keOXKEWbNmkZWVxU033UT58uUdPkZBFhcXx/z58wkICKB37974+flddpsFCxawYcMGKleu\nzM0331xkl3l1O/lZkP5KHlgz8tcDqcCLl/j9ZqDMRa93ACGXeF/uq+u7UHZ2ttwz+F6pXKO2dL1j\nsJSrVEVGPfFkrtucOnVK6jdoKE1atZPOfe6W4NAwmTdvnosSO9bx48clMDhEajVpIR1u6y9ePr7y\n5JO57/+2bdskLCJC2t50u8R0v1UiSkXKzp07c93myy+/lJCwcOnSd5A0vLaVNItuIWfOnHHkruQq\nOVmkd2+R2rVFNm68us+aPXu2hAT4SvfaEdK0Qpg0vqaepKWlOSboeSkpKRLi7yu1wr2lS7Ug8SlZ\nTB568MFct9m1a5eUCguR9tUjpGONCAkPDnJKQ6ORjw+X8uGB0qNOhFQpFSz3DrrL4WPs2bNHIsND\npV31cOlUI0LCggIlLi7O4eMUVD/99JME+/vKjbUjJLpyuNSpUU1SU1Nz3WbsmKckKsz6c6kWGSx3\n9uldKJszuTMc3TjG0YwxAcAcYJiIbLno55uBTiJy8PzrnUAzEUn61/biqqyX88cff9CmfQcmzFqE\nl48vJ5NP8HjP69keH59jQ4wpU6YwfdZcHn7lfYwx/LFqGV+8+gw7t29zcfqr179/f9Zt28XoqTMx\nxrB17QpeffRuTp9MzXGb3n37Uiy8Et0GWjPq57z/Bl5px5n+0Yc5blOmbBRDXpxK1XoNEREmDruD\nR+8flKej0au1fLl1Gr5zZ3j1Vbjag+zyZSJ5oI4ntcN9EBFeWHGc+55+iUGDBjkmMDBkyBBiZ03n\npfblMcaw4/gZnv51P2cysnLc5u4Bd3J2w3x617EmXc6JP8GpStF8NXuOw3IdPHiQWtWrMrlTFAGe\nxTmTkc2wBQdZsmK1Q1v43jvoblJXf0/fuqEAzNt2gmNlGvPtvP85bIyCrE6NqvQqfY7GZayj9VdW\nHePmYU/xyCOXXiwmKSmJ8lFlmHJDOYK8SnAuM5uHFx7mf7/E0rBhQ1dGV7kokNfcLyYiqcaYRcAN\nwJaLfpUAlAMOnr8uH/Dvwv6XcePGXXgeExNDTEyM0/LmJikpibDIMnj5WF3d/IOCCQwOJSkpKcfi\nnpSURGSFqhdOeZWpVJWkpOMuy+xIiYmJlK9a88K+lK1cjYzLLD97/HgS1zTrcOF16YpV2bE49y82\nySdOUKaSdY+ZMYbICpU5fty5/86ysuDFF+Gtt2DaNOjZ0zGfeyIllagA63S3MYYyPsUcvi9Hjhyh\nfKDHhT+XqABP0rOyc93mWOJRavr//b+BKP+SLD161KG5Tpw4QbCvFwGexQHwLlmMCH9vh+//8cSj\nVPS7aF8CPNh2LNGhYxRkSUkniKoRdOF1GW84lsv+Jycn4+/tSZCX9e/Ms0QxSgV4Of3vmMpdbGws\nsbGxV/05Tl3ExhgTZowJPP/cG+gAxP/rbfOAAeef3wr8mtPnjRs37sLDrsIOUK9ePY4fOsDS/31D\n2qmTLPjyY4ojubYkbdeuHb//8A0749ZzKuUEX731Ih06dMjx/QVZ7969if3uK3ZuXs+p1GQ+fXU8\nZcqWzXWbzjd04vuP3uLYoQMkHtzPD9On0OWGTrlu065DB2a+8TynT6awfeMaVv3yP9q0aePIXfmH\n/fv/XhN+7VrHFXaA9u3a8mlcMqfSs9h27AxLE9Jo27at4wYA+vbty+K9J9mSmMap9Cw+WH+EyLDc\nb4Ps3K0Hc/5M48ipdBJPZzB752k639jDobmqVq2KlPTihx3JpGVksWhPKsfPZlGvXj2HjtO5Ww/m\n7jrDoZPpHEvLYNaO03Tu1t2hYxRknTp1YsaWVFLPZfFn0lkWJZylQ4eOOb6/fPny+PoH8t12689l\nyd5UElLO6VG7zWJiYv5R6/ItP+fy8/oA6gHrgA3AJuCp8z8fD9x4/rkn8BXWtfYVQMUcPsuBVzGu\n3rp166ROvfri7e0jjZs2k/j4+MtuM2PGDClTNkp8/fzllltvk5SUFBckdY6HHn5YvHx8pXiJklK+\nYiXZu3dvru/PysqSkaOekKDgYAkKCZGnnnr6stf2kpKSpFuPm8TH10+iypWX2bNnO3IX/mH2bJGI\nCJEXXhDJzHT85ycnJ0uPrp3F19tLypYKly+//NLxg4jIE088IT4li0uJYkbKRoRedl5Ddna2jHlq\ntAQH+EmQv588/tijkpWV5fBc8fHx0qh+XfH29JC6NavL+vXrHT5Gdna2jHtmjIQE+kugn688+tCD\nkumMP8wC6tSpU9K7183i5+MtkeGh8tGHH152m507d0qzRg3E29NDalWrIqtWrXJBUnUlKOjX3K9W\nQbrmrtxHWho8+igsWGA1fNG1NZRSBYk2jlHqCm3cCE2awOnT1r3rWtiVUu5Ci7uLnTlzhmPHjmk/\nYxuJwJtvWivMPfmkteKcE1bctM25c+dITEy8ov/GUlNTSUlJcWKq/Dl37hxHjx69on05efLkFe2L\niHDs2DHOnDmTn4h5lp99uVKu2hdV8Glxd6H/e/FFQkJCqVylKo2bNuPQoUN2RypyEhOhWzerg9uK\nFXDnnXYncqw3J71OcGAAVStVoF6tGuzduzfX92dkZHBHn9uJjAijdKkIbunRjXPnzrkobe4mv/02\nIUGBVKtckTo1qrFr165c35+ZmcmAO/pSKjyU0qUiuOnGLpctcocPH6ZZowZULl+OkKBAnhs/zoF7\n8LdpU6de2Jda1aqwc+dOh49x9OhRmjdpeGFfxj1TOJsyKcfQa+4u8tNPP3HPfUMY/d7XBIWV4uvJ\nL3Nq/w4WzHdsr3WVswULYOBA6N8fnn0WSpa0O5FjLVu2jFu6deb56yMI9ynB7Phkthcvw4o163Lc\n5oXnn+Pb995gVPNQjIHXVifRutdAXnx5oguT/9eKFSvo0bkjz18fToRvSeZsS2ZzdgRrNmzKcZsJ\nL73Il1Ne5YnoUIobeH1NEtE97uCV117PcZsb2rfF/8gf3FE3mBNns3hmaSLTpn9Oly5dHLYva9as\noXP7Njx/fQSl/T34btsJ1pwLZkPcVoeNAdCtcydK7t9A/3rBpJ7N4pllibzx3nR69HDs3Q/KtfSa\newG3atUqmra/keDwSIwxdOpzN2vWrLr8huqqpafDiBFw113w6afWfezuVtjB+m+seWkfInxLYoyh\nW7VA1m3MuRgCLF+6mHblPPEsUQyP4sXoUN6L5cuWuChxztasWUOTMj6U8rPu27+xWhAbNv+R6ynt\nFUsX07acJ14lilGyeDE6VvBmxWX2Zc3atXStGoAxhhDvErSILMnKlSudsC9+lPa32gd2rRbE5i3b\nyMzMdOg4q1avpmsVf4oZQ5B3Ca6NLMnKFSscOoYqPLS4u0j58uX5c/Nass7/hd66biVR5YrWutd2\n2L4dWrSw/rlhAzj41vICpUKFCmxPziDj/MI1fxxNo2xkqVy3qVi5CltP/F1ktiRlULFSZafmzIvy\n5cuzIyn9wiI8WxLTKF0qPNd1zytWrsrWpIwLXwC2HE+/7L5ElS1D3FGr/UVWtrA9RahQoYKD9sJi\n7cs5zmVa+7L12BnCQ0MoUcKxa4iVj4rij4v3JVWoULGiQ8dQhUh+7p+z40EBu8/9SmVkZMgNXbpK\npeq1pFlMBwkNC5cVK1bYHcttZWeLfPihSGioyNtvW6/dXVZWlvS6qbtUiAiSa6tGSnCAnyxatCjX\nbRITE6VGlUpSv3y4NKwQIZXLR8nBgwddEzgXWVlZ0rvXLVIh/O99WbBgQa7bHD9+XGpXryr1y4dL\no4oRUiGqjOzfvz/XbVatWiWhQYHSokqkVIkMlo7t2kh6erojd0Wys7Pljj63S7mwQGlZzdqXn3/+\n2aFjiIisWbNGwoIDJbpKKakaGSLtY1o5fF+U66H3uRd8WVlZLF68mOTkZKKjo/PUFU1dueRkuP9+\niIuDL76AunXtTuQ6IsKSJUs4duwYzZo1y1OHt7S0NBYtWoSI0Lp1a/z9/V2Q9PJEhKVLl5KYmEjT\npk0pV67cZbdJS0sjNjaWrKwsWrduTUAeboM4fPgwy5cvJzAwkNatW1O8eHFHxP8HEWHZsmUcOXKE\nJk2aOPzswF+OHDnC8uXL8ff3JyYmxin7olwrv9fctbgrt/L779CvH3TtChMnXn3DF6WUslOBbxyj\nlDNlZcELL8DkyfDuu6AThJVSRZlOqFOF3r590KYN/PYbrFunhf1Kvfjii5SPDKNcqTDGjCnc90ZP\nnDiR8pHhlCsVypNPPml3nKuyaNEienS5gS4d2jF79my741ywb98++vftQ/vW1/H8s+MdPutfOYYW\nd1WozZ4NTZtCly4wfz5cpjmd+peJEyfy/NinubmSB7dV8WDSyy8W2gI/adIkxj71BD0rluD2Kl5M\nfm0io0aNsjtWvixdupReN3WjQtJGaqfFM3TwXXz55Zd2x+L48eNc27wpGX/8Sgv2Mvv9Nxly72C7\nY6lL0GvuqlA6fdpq+LJwodXwpXlzuxMVTpXLlaZLGehYxeoDvnhvKjO3n+FAYpLNya5c1QpRdIjI\npHO1YAB+35/KJ1tPc/BYss3JrtydfXvjsz2WrtWtfVmZcJKlmWX57XfH3oN/pWbMmMGUsY8xqpnV\nSjgtI4v+c3aTduaMw2/tUxZdxEYVGRs2WA1fzp6F9eu1sF8Ng4GLvjOLSK73khdk/86dXYiPBYwx\nZF90MJMt/90/O7jTv2N3p8VdFRoiMGkSdOgATz8Nn3ziXg1f7PDg8JF8uOEoC/5M5tddKbyz5giD\nhz5kd6x8efjxUXy8IZH5O5P5dXcKU1cf4a77htodK1+GDHuIb3am8dPOEyzancKHcSk8NHyk3bHo\n3Lkz+88U47PNSSzff5IJK48z6O6BetReAOlpeVUoHD1qrQt//DjMnAmV7V9EzW289tprvPnqBBC4\n54EHefrpp+2OlG9vvPEGk15+kezsbO6+fyhjx461O1K+LV26lEmvTCAzI4NB9w+lW7dudkcC4MCB\nA4x9ejQH9u+nTfuODB8xQu+ndyK9z125rfnzrXXhBwyA8ePdc114pZS6FL3PXbmd9HQYPRq+/NLq\nud6mjd2JlFKqcNDirgqk7duhTx8oV86aQBca6ryxsrOzWbhwIUlJSbRo0YLy5QtvQ5/s7Gx+/fVX\njh07RnR0NBWd1Dhk9erVfPbZZ0RGRjJ8+HA8PDwcPkZ2djaxsbFWn/LmzalUqZLDx1DKXelpeVWg\niMBHH8GoUVbP9fvvB2dOEs7KyuKmG7sQv2E1Zfw92XzkFN/MnUdMTIzzBnWSrKwsbunRjc1rVhAV\nYO3L19/MoV27dg4dZ+rUqTz60DDqhHtz+FQG2R6+7E44iJeXl8PGyM7O5vZberJu+RLKBngSd+QU\nM7+aRadOnRw2hlKFgV5zV4VecjLcdx9s3WpNmqtTx/ljzpw5k/8b+SDjW4ZRophh7cFTzNgDO/fs\nd/7gDjZr1iyeefg+nr8+nBLFDBsOnebDHZnsSTjo0HECfDwZ1iSc6Ch/srKFJ37ZS6ued/Dee+85\nbH3T45oAABr/SURBVIw5c+bwxAODeOH6cEoWN2w+cpqpW86RcOiIw8ZQqjDQ+9xVobZ0KTRoAOHh\nsHKlawo7QEJCAlUCi1OimPV3p1a4NwcPH3XN4A6WkJBA1aAS/9yXI4kOH+dsega1wqyOPMWLGeqE\n+7Bnzx6HjpGQkECV4JKULG7tS80wHw4fPYZ+wVcqb7S4K1tlZloz4Hv1grfegrffdm0nt+joaFYc\nPMORU+mICHO2p9CsSWPXBXCg5s2bs/JgGocv7EsyTRs1cPg4YcFBfLM1CREh8XQGi/ak0rFjR4eO\n0axZM1YfPM3Bk9a+zN2eTOMG9QvEQi5KFQr5aQJvx8OKqtzJnj0i110n0q6dyIED9uV4c9Lr4u3p\nId6eHtKofh1JSEiwL8xVmvz22xf25Zq6tWTfvn0OH2Pjxo0S7O8jJYoZKW6Q7jd2dfgYIiLvvDNN\nfLw8xdvTQ+rXril79uxxyjhKFWTna98V10y95q5s8fXXMHQoDB8OI0ZAMZvPIWVmZnL69GkCAwPt\nDeIArtqXhIQEQkJC8PHxcdoY7vTnolR+6IQ6VSicPg0PPwyxsdakuaZN7U6klFIFl06oUwXe+vXQ\nuDFkZFjPtbArpZRzaHFXTpedDa+/Dh07wjPPwPTp4O9vd6qCb/PmzdStWR2PkiWoXb0K69evd/gY\n2dnZPDlyBP6+Pvj5ePPIg8PIysr6//buPMqK6tz7+PehB0SaGQEVgSgoU7AZZBDFDslVMASyjAaV\nG+chvr5B5b4sHHIjRq8mRpMoGnFk8JUQQ8SRySgtKGIDDSIoijKDTDILPdHP/aMKbYGmoe0zVf8+\na/XinKpdp57tbvs5tWvX3kc8ZseOHQz8aT9qZmbQpFFDxo8bV+Vxicj3o255ialNm4IFX7ZvD9Zd\n14IvR2fv3r20ObUVv2iVzrkt6vD+ut1MWF7I8hWrqFOF34weHzWKUQ/czYjuDalhxsPzt3H5TcO4\n/c67yj3m4p8PZPcnc7mmUwO+3F3E/8zdyitTpnP22WdXWVwiElC3vCSdadOgc+egK372bCX2Y/HZ\nZ59Ry/bzk1PrUTO9Bjmt6tGgZg0+/vjjKj3PtCmvMfDUWjQ6PoMGtdL5+WnHM2PqG0c85q23ZzKk\nfT1qZdTg1IbHcV7zWuTm5lZpXCLy/Si5S5UrLIRhw+CGG4JBc/fdp5XcjlXjxo3ZunsfuwuDLvI9\nRfvZvGsfjap4kv0mTZuxZnfJN+/X7CrmhCZNj3hMowb1Wb2zEAgepV37tdO4ceMqjUtEvh91y0uV\n+vTTYMGXVq3gmWegYcNER5S6Rgz/f0wc+wxnNjmOxVsK+fngIfx11GNVeo7Vq1fTq3s3zqhXgzSD\nxVuLeW/uB7Rp06bcY6ZMmcKvLhtMr+a12fj1fmo0PJl33nufWvGcfUikmtCjcJJQ7vDss3DHHcGV\n+g03xHbBl+pixowZfPzxx7Rt25YLLrggJjO0bd68mcmTJ+PuDBw4kJNOOqnCY5YsWcLMmTOpX78+\nl1xySZUuGiMi30rK5G5mzYHxQFOgFHja3R89qMx5wCvAinDTS+5+32E+S8k9SW3fHiz4smwZTJwI\n7dsnOiIRkWhI1gF1JcAwd+8A9AJuNrO2hyk3y927hD+HJHZJXrNnBwu+NGsGeXlK7CIiySA9lh/u\n7huBjeHrPWb2CXAysOygourATTElJXDvvfDkk8G99QEDEh1RoKioiEcfeYSlixfRoVM2t9x6Kxkp\nOpqvuLiYUY8+wkeLFtL+h2dyy623kpmZecRjtmzZwkN/epAtmzZywYUDGDx4cJyirT5KSkp4/PHH\nWDR/Hqe3a89tw/5LtyUk6cTtnruZtQJygY7uvqfM9vOAScA6YAMw3N0Ped5H3fLJY/VqGDIkWL1t\n/Hg48cRERxRwd37W/wI2L8unW5N05m0u4cT23Xjljakpt5qYu3PRwAGsXfwB3Zums2BLCY3aZPP6\ntBnUKGci/h07dtC5U0c61C7klKw0pq4u4KZhIxhx+x1xjj663J3LB1/Csrm59GqWzqKt+zmu+Rm8\nOfMd0tLSEh2eRFCydssDYGZZBAn8lrKJPbQAaOnunYHHgJfjEZNUzosvBtPGDhoE06cnT2IHWLZs\nGfPz5jKiR2P6tW7A7T0ak/f+HD777LNEh3bMPv/8c96bPYs7ejaiX+sGjOjRmEUL5h3xOfdJkybR\nPLOIGzo3pn+bBtzRoxF/uP/+OEYdfevXr2fa1CncGbbL8O6N+GLZUvLz8xMdmsh3xLRbHsDM0gkS\n+/Pu/srB+8sme3efamZ/M7OG7r7t4LIjR4785nVOTg45OTkxiVkOtWdPsODLrFkwZQp065boiA5V\nUFBArcx00sOvrOk14LjMNAoKChIbWCUUFhaGdQm+sKcZHJ+ZTmFh4RGPqZ3x7Rf8rMw0ioqLYx5r\ndVJQUEDN9HQy08J2qWEVtovIscjNza2SSaFi3i1vZuOBre4+rJz9Td19U/i6O/Ciu7c6TDl1yydI\nfn7w7HqvXjBqVPLOC19UVESXTh1pm7GTXifVYs6GfXxWUp/8xUtS7r57cXEx3bI7cZpto/fJtfjg\nywKWFGSx8KOl1KxZ87DHrFq1iq7ZnRjSNosW9TJ58dM9dDyvP8+Nez7O0UfX/v376dmtC80KN5Jz\nSi0WbCwgb2cmiz9eFtOlb6X6SspueTPrDQwB+prZQjPLN7N+Znajmd0QFrvYzJaY2ULgr4BGACWJ\n0lJ4+GHo1w/uuQfGjk3exA6QmZnJv3NnUeP0XoxdlUHa6b34d+6slEvsABkZGbw58x0y253D2FUZ\neOuevPXO7HITO0CrVq2Y8dZMlmaeyvNrMulz0X/yxFPPxDHq6EtLS2Pav9+mXqfzGLsqg30tupL7\n7hwldkk6msRGDmvjxmDBl1274IUX4Ac/SHREIiLVT1JeuUtqmjIlWPCle/fgHrsSu4hIaon5gDpJ\nHYWFMGIEvPQS/OMf0KdP1X5+cXExixYtwszIzs4mPV2/fhXJz89nzpw59OzZk27JOIpRRJKS/roK\nAJ98EgyaO+00WLSo6hd82bFjBz/J6cO2jetwd5o0b8WbM9+hbt26VXuiCLn5ppt49uknaVYnk417\nirj8P6/kuTFjEh2WiKQA3XOv5tyDGebuvBPuvx+uuy42C77c/OsbWfnOy9yY3QAH/pa/jY79LuXP\njzxa4bHV0fLly+nY7gwevqAVLerVZP2uIm6btpK58/PJzs5OdHgiEieVveeuK/dqbNu2YPW2zz8P\n7q23axe7cy37eAm9m2RiZhjQtUkmi5Yuid0JU1xeXh4Na2XQol4wOv7kupmcUDuTvLw8JXcRqZAG\n1FVTs2YFC740bw5z58Y2sQOc2bkL724oZH+ps7/UmfNlEWd27hrbk6aws88+m237iln+1T4AVmwv\nYPPXRfTu3TvBkYlIKlC3fDVTUgK//z08/XSw/vqFF8bnvHv27GFAv/NZtnQJjvPDMzvz6pRpej74\nCO644w7+/Kc/Uq9mOjsLS/i/v7mFh//yl0SHJSJxlJTruVclJffvb9UquPxyyMqCcePiPy98aWkp\nX3zxBQCtW7dOucVcEmHNmjXk5eVx1lln0bJly0SHIyJxpuQuRzRxIgwdGjzqdtttUM7CYiIikkQ0\noE4Oa88e+M1v4L33YOpU6Krb3CIikafrtyTm7vzpoYdoc0Y7zmjXgb898cQxHT9/PnTpElyl5+en\nXmLfu3cvN1x7Na2an0TXTh2rZKWkVLJ27Vr6/aQvLU5qxn/8qA8rV65MdEgCbNiwgQH9zqfFSc3o\n26c3y5cvT3RIIodQt3wSe2L0aB78yyNc97uH2V9SwlN338qfHvgfLrvssiMeV1oKf/4zPPggPPYY\n/PKXcQq4ig0ZfAlr5udyWbs6rN1ZxJOLdzLng3m0bds20aHFXFFRET9sdwZn1d3HOc1rM3f9XmZt\nTWPpp8upVatWosOrtkpKSuj8ww60z9jBj1pmMf/LvczYAEs//Yw6ybyqkqQszS0fQS9O+hcX/58R\nnNq+E206dWHQ9bfx4qR/HfGYL78MVnGbPBny8lI3sQO88trr/Dq7Ac3r1qTXKXU4++TjmT59eqLD\niotly5ZRtGcnv2zXgJPqZHJR2/qklxSwZInmBkiklStX8tXmjVzeIWiXgafXp05aCQsXLkx0aCLf\noeSexOpkZbF988Zv3m/fvPGIVwdvvBF0w/fqBe+8A61axSHIGDr+uOPYtq/km/fbCp2srKwERhQ/\nWVlZ7C4oorCkFICi/aXs2ldUbeqfrGrXrs3XhcXsC9uleL+zc1+x2kWSjrrlk9i8efO4oF9/+gy6\nlNLS/bz3xiRm5ebSoUOH75QrKAhGwU+eHCzPeu65CQq4io0e/QT33DmC/zjlONbvddZ7HeYt/LBa\ndH+6O1cOuZz8WTPo2jiNhV/tp13385g46V96hDDBbrj2GmZPfZnuTdL5aHspp3Q8i5dfn6J2kZjQ\no3ARtXTpUiZMmECNGjW44ooraNOmzXf2f/IJXHoptGkTTEzToEGCAo2RGTNmMGP6NBqf0ISbbrqJ\nevXqJTqkuCktLWXcuHF8tPhD2nfoyNVXX01aWlqiw6r23J3nn3+eRQvzOf2Mtlx33XVa4VBiRsm9\nmnEPkvldd8EDD8C118ZmwRcREUkcPedejWzbBtdfDytWwOzZUA0Gj4uIyDFQck8xublwxRVw8cUw\nYQLUrJnoiERiY9++fYwdO5ZNmzbRt29f+vTpE5PzzJs3j6lTp1K/fn2uuuoq6tatG5PziMSTuuVT\nRHEx3HNPsNjLmDHB424iUVVQUMC5vXqQvmMDp2QZs9YVcN+DD3Pd9ddX6XkmT57M9VddQc4ptdhS\n4GyiDnn5i5TgJWnonnuErVgBQ4ZAvXrBgi9NmyY6IpHYmjBhAg/dcQt3n90IM2PNzkJ+9+5Wtu3c\nVaXnad2qBde0Njo2CVYnfDhvK78Y+t8MHTq0Ss8jUlmaxCai/v536NEjmIxmyhQldqkeduzYQdPa\nad88Xta0dgZ79u6lqr/g79y1i6a1M7553+Q4Y/v27VV6DpFEUHJPUrt3w5VXwsiRMH26VnKT6qVv\n3758sG4P+Rv28NXeYp75cDvn/7hvlT9LPmDAAMYu2cnWvcUs3byXmWv30b9//yo9h0giKF0koXnz\ngpnmMjKCBV+6dEl0RCLx1bZtW1586WUmrstg+Dtbqdu+F/9/4otVfp7HRz9F67PPZ8SsrTyzvJQn\nnxtL9+7dq/w8IvGme+5JpLQUHnoo+EnlBV9ERKRq6Dn3FPfll8Ejbvv2BVfuLVsmOiIREUlV6pZP\nAq+/Dp07wznnBM+xK7GLiMj3oSv3BCoogOHD4bXXYNKkILmLiIh8X7pyT5ClS6F7d9i0CRYuVGIX\nEZGqo+QeZ+4wejTk5MCtt8I//hG9ldxERCSx1C0fR199BdddB6tXw7vvwhlnJDoiERGJIl25x0lu\nLmRnw2mnwfvvK7GLiEjs6Mo9xoqLg1nmxoyB557Tgi8iIhJ7Mb1yN7PmZva2mS01s4/M7LCrMZjZ\no2a23MwWmVl2LGOKpxUr4NxzgwFzCxcqsYuISHzEulu+BBjm7h2AXsDNZta2bAEz6w+c5u5tgBuB\n0TGOKS5eeCFY8OXSS4Pn2LXgi4iIxEtMk7u7b3T3ReHrPcAnwMkHFRsEjA/LfADUM7OUTYW7dgUz\nzd17L7z5ZjAiXgu+VE5JSQn33jOS3t27Muin/ViyZEmiQxIRSQlxSztm1grIBj44aNfJwNoy79dz\n6BeAlJCXFyzyUrMmLFgQDKCTyht261AmPTOKfllbaLrpQ3507jmsWbMm0WGJiCS9uCR3M8sCJgG3\nhFfwkVJaCn/4AwwYAA88AE8/DbVrJzqq1Dd27Dhu69aQM5vVZsDp9enStCavvvpqosMSEUl6MR8t\nb2bpBIn9eXd/5TBF1gOnlHnfPNx2iJEjR37zOicnh5ycnCqLs7I2bIBf/SoYFT9/PrRokeiIoiOt\nRg2K95d+87641ElLS0tgRCIisZWbm0tubu73/pyYL/lqZuOBre4+rJz9FwI3u/tPzawn8Fd373mY\nckm35OvOndC+Pfz613DnnaC8U7V+P/Juxo9+lJ/9oBbr9uxnzhZn0UdLOeGEExIdmohIXFR2ydeY\nJncz6w3MAj4CPPy5E2gJuLs/FZZ7DOgHfA1c7e75h/mspEvuAOvWQfPmiY4imtydMc89x7Q3XqVx\nk6bc+dvf0Vz/sUWkGknK5F6VkjW5i4iIxEplk7se0hIREYkYJXcREZGIUXIXERGJGCV3ERGRiFFy\nFxERiRgldxERkYhRchcREYkYJXcREZGIUXIXERGJGCV3ERGRiFFyFxERiRgldxERkYhRchcREYkY\nJXcREZGIUXIXERGJGCV3ERGRiFFyFxERiRgldxERkYhRchcREYkYJXcREZGIUXIXERGJGCV3ERGR\niFFyFxERiRgldxERkYhRchcREYkYJXcREZGIUXIXERGJGCV3ERGRiFFyFxERiRgldxERkYhRchcR\nEYkYJXcREZGIUXIXERGJmJgmdzN71sw2mdnicvafZ2Y7zCw//PltLOMRERGpDmJ95T4GuKCCMrPc\nvUv4c1+M40lZubm5iQ4hoVT/3ESHkDDVue6g+lf3+ldWTJO7u78LbK+gmMUyhqio7r/gqn9uokNI\nmOpcd1D9q3v9KysZ7rn3NLOFZvaGmbVPdDAiIiKpLj3B518AtHT3vWbWH3gZOD3BMYmIiKQ0c/fY\nnsCsJfCau3c6irIrga7uvu0w+2IbqIiISBJy92O+fR2PK3ejnPvqZtbU3TeFr7sTfNk4JLFD5Son\nIiJSHcU0uZvZBCAHaGRma4C7gUzA3f0p4GIzuwkoBvYBg2MZj4iISHUQ8255ERERia9kGC1/CDOr\nEU5q8+ph9mWa2UQzW25m75tZi0TEGCsV1P1KM9tcZtKfaxIRYyyZ2Soz+zB8giKvnDKPhu2/yMyy\n4x1jrFRU96hP+mRm9czsn2b2iZktNbMehykTybaHiusf5fY3s9PD3/v88N+dZjb0MOUi1/5HU/fK\ntH2iR8uX5xbgY6DuYfZdC2xz9zZmNhh4ELg0nsHF2JHqDjDR3Q/5pY+QUiDH3Q87P0L4VMVpYfv3\nAEYDPeMZYAwdse6hWe4+MF4BxdkjwBR3v8TM0oHjy+6MeNtDBfUPRbL93f0zoDMEFzjAOmBy2TJR\nbf+jqXvomNo+6a7czaw5cCHwTDlFBgHjwteTgB/HI654OIq6Q/Qn/TGO/Hs5CBgP4O4fAPXMrGk8\nAouDiup+oEzkmFld4Fx3HwPg7iXuvuugYpFt+6OsP0S0/Q/yE+ALd1970PbItn8Z5dUdjrHtky65\nA38BhgPlDQY4GVgL4O77gR1m1jBOscVaRXUHuCjsknox/DIQNQ5MN7N5Znb9YfZ/0/6h9eG2KKio\n7hDdSZ9+AGw1szFht+NTZlbroDJRbvujqT9Et/3LGgz8/TDbo9z+B5RXdzjGtk+q5G5mPwU2ufsi\njvAI3cGHxTaq+DjKur8KtHL3bODffNuDESW93b0bQQ/GzWZ2TqIDiqOK6n5g0qfOwGMEkz5FRTrQ\nBXjc3bsAe4HbExtSXB1N/aPc/gCYWQYwEPhnomOJtwrqfsxtn1TJHegNDDSzFQTfXn5kZuMPKrMO\nOAXAzNKAuuU9G59iKqy7u2939+Lw7TNA1zjHGHPu/mX47xaC+07dDyqynrD9Q83DbSmvorq7+x53\n3xu+ngpkRKjXah2w1t3nh+8nESS7siLb9hxF/SPe/gf0BxaE/w8cLMrtD0eoe2XaPqmSu7vf6e4t\n3P1UgkFyb7v7FQcVew24Mnx9CfB2PGOMlaOpu5k1K/N2EMHAu8gws+PNLCt8XRs4H1hyULFXgSvC\nMj2BHQcmQkplR1P3svcXrYJJn1JN2IZrzezA9NM/5tDf70i2PRxd/aPc/mVcRvnd0pFt/1C5da9M\n2yfraPnvMLN7gHnu/jrwLPC8mS0HviJaI+UPcVDdh5rZQIJJf7YBVyUythhoCky2YKrhdOAFd59h\nZjcSTnzk7lPM7EIz+xz4Grg6kQFXoQrrTvQnfRoKvBB2T64Arq4mbX/AEetPxNvfzI4nGFB2Q5lt\n1aL9K6o7lWh7TWIjIiISMUnVLS8iIiLfn5K7iIhIxCi5i4iIRIySu4iISMQouYuIiESMkruIiEjE\nKLmLpAgzu8vMlliwLGy+mZ1VxZ9/npm9drTbq+B8g8ysbZn3M83s4FnpRKQSUmISG5HqLpyR60Ig\n291LwqknM2NwqvImvojFhBg/B14HlsXgs0WqNV25i6SGE4Gt7l4C4O7b3H0jgJl1MbPccDW5qQem\nqgyvhP8ariS12My6hdvPMrM5ZrbAzN41szZHG0Q4Te6zZjY3PP5n4fYrzexf4fk/NbM/ljnm2nDb\n3HC1s1Fm1otgkYwHw16IU8PivzSzD8xsmZn1ror/cCLVkZK7SGqYAbQIk97jZtYHwMzSgVHAL9z9\nLGAMcH+Z42qFK0ndHO4D+AQ4x927AncDDxxDHHcBb7l7T6Av8FCZpUnPJFjvoRMw2MxONrMTgd8S\nLILTG2hLMKXm+wRzhQ939y7uviL8jDR37wHcBow8hrhEpAx1y4ukAHf/OrwffS5BUp1oZrcTLAXZ\nEXjTzIzgC/uGMof+PTx+tpnVMbO6QF1gfHjFfmAu+6N1PvAzMxsevs8EWoSv33L3PQBmthRoCZwA\n5Lr7znD7P4Ej9RS8FP67IDxeRCpByV0kRXiwEMQsYJaZfUSwQlY+sMTdy+vCPvheuQP3Eqw6eJGZ\ntQRmHkMYRtBLsPw7G4MxAYVlNpXy7d8XO4bPP/AZ+9HfJ5FKU7e8SAows9PNrHWZTdnAauBT4IQw\nuWJm6WbWvky5weH2c4Cd7r4bqMe362Af68pa0wlWLzsQV3YF5ecBfcysXngL4Rdl9u0m6EUoz7F8\nKRCRMpTcRVJDFjAufBRuEdAOGOnuxcDFwB/D7QuBXmWOKzCzfOBvwDXhtgeBP5jZAo79b8C9QEY4\nQG8J8PtyyjmAu28gGAOQB8wGVgI7wzITgeHhwLxTOXwvg4hUgpZ8FYkoM5sJ/Je75yc4jtrhmIE0\nYDLwrLu/ksiYRKJOV+4i0ZUs39xHmtlC4CNghRK7SOzpyl1ERCRidOUuIiISMUruIiIiEaPkLiIi\nEjFK7iIiIhGj5C4iIhIxSu4iIiIR87+4hiYzcuU4oAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe54d577cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(2, figsize=(8, 6))\n",
    "plt.clf()\n",
    "\n",
    "# Plot the training points\n",
    "plt.scatter(X[:, 0], X[:, 1], c=Y, cmap=plt.cm.Paired)\n",
    "plt.plot(x1,y1)\n",
    "plt.xlabel('Sepal length')\n",
    "plt.ylabel('Sepal width')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Create perceptron\n",
    "threshold = 1\n",
    "\n",
    "# Get weights using the line equations from above\n",
    "w2 = threshold / c\n",
    "w1 = -1 * w2 * m\n",
    "\n",
    "# Perceptron calculation\n",
    "plant_class = ((X[:,0] * w1 + X[:,1] * w2) > threshold) + 0\n",
    "\n",
    "# Compare\n",
    "correctly_classified = np.all(plant_class == Y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correctly_classified"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A real perceptron"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0]: 0.227399965745 -> 0\n",
      "[0 1]: 0.704595758278 -> 1\n",
      "[1 0]: 0.325934901387 -> 1\n",
      "[1 1]: 0.80313069392 -> 1\n"
     ]
    }
   ],
   "source": [
    "from random import choice\n",
    "from numpy import array, dot, random\n",
    "\n",
    "def unit_step(x,theta):\n",
    "    ''' This is our activation function. '''\n",
    "    fx = 0\n",
    "    if x < theta:\n",
    "        fx= 0\n",
    "    else:\n",
    "        fx= 1\n",
    "    return fx\n",
    "\n",
    "training_data = [\n",
    "    array([0,0,1]),\n",
    "    array([0,1,1]),\n",
    "    array([1,0,1]),\n",
    "    array([1,1,1]),\n",
    "]\n",
    "\n",
    "target = [0,1,1,1]\n",
    "\n",
    "w = random.rand(3) # Initializing weights randomly.\n",
    "errors = []\n",
    "eta = 0.2 # Learning rate\n",
    "n = 100 # Number of iterations\n",
    "theta = 0.25 # Threshold\n",
    "\n",
    "for i in range(n):\n",
    "    x, expected = choice(list(zip(training_data,target)))\n",
    "    result = dot(w, x)\n",
    "    error = expected - unit_step(result,theta)\n",
    "    errors.append(error)\n",
    "    w += eta * error * x\n",
    "\n",
    "for x, _ in zip(training_data,target):\n",
    "    result = dot(x, w)\n",
    "    print(\"{}: {} -> {}\".format(x[:2], result, unit_step(result,theta)))\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe54d02e9d0>]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEACAYAAAC+gnFaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAknqnEEgSZ0zCCSpcwaBJHXOIJCkzv0//0b0Eew3MQoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe54d11d890>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from pylab import plot, ylim\n",
    "ylim([-1,1])\n",
    "plot(errors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0, 1)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAEACAYAAABWLgY0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAE6VJREFUeJzt3XuM3WWdx/H3ty0X5RpAQNuCQLnUFgpDgvyxK8dglkJU\n2Bi0LK6s6xp10zVxQ8RNTCjxH90/8MJqdDdg1mVNXTdmZV2JmMWjIS7YzLQUSksrQm80QIHacG1p\nv/vH70znzKUzpzPnMvPM+5Wc9Jwzz/zON0+mn3nm+T2/3xOZiSSpTHN6XYAkqXMMeUkqmCEvSQUz\n5CWpYIa8JBXMkJekgk0Y8hFxd0Q8FxHrx2nzrYjYEhHrIuLS9pYoSZqsVkby3weuOdwXI+Ja4LzM\nPB/4DPDdNtUmSZqiCUM+Mx8CXh6nyfXADxptHwFOiogz2lOeJGkq2jEnPx/Y3vR6Z+M9SVKPeeJV\nkgo2rw3H2AksbHq9oPHeKBHhjXIkaRIyMybzfa2O5KPxGMt9wCcAIuJKYE9mPne4A2Wmj0xuv/32\nntcwXR72hX1hX4z/mIoJR/IR8UOgBpwaEduA24Gjq7zOf87Mn0fEdRHxe+BV4JNTqkiS1DYThnxm\n/kULbVa2pxxJUjt54rVHarVar0uYNuyLIfbFEPuiPWKq8z1H9GER2c3Pk6QSRATZ4ROvkqQZyJCX\npIIZ8pJUMENekgpmyEtSwQx5SSqYIS9JBTPkJalghrwkFcyQl6SCGfKSVDBDXpIKZshLUsEMeUkq\nmCEvSQUz5CWpYIa8JBXMkJekghnyklQwQ16SCmbIS1LBDHlJKpghL0kFM+QlqWCGvCQVzJCXpIIZ\n8pJUMENekgpmyEtSwQx5SSqYIS9JBTPkJalghrwkFcyQl6SCtRTyEbE8IjZFxOaIuG2Mry+MiAcj\nYiAi1kXEte0vVZJ0pCIzx28QMQfYDFwNPAusAVZk5qamNt8DBjLzexGxGPh5Zp4zxrFyos+TJA0X\nEWRmTOZ7WxnJXwFsycytmbkfWA1cP6LNQeDExvOTgZ2TKUaS1F7zWmgzH9je9HoHVfA3uwN4ICI+\nD7wd+EB7ypMkTUUrId+Km4DvZ+bXI+JK4F5gyVgNV61adeh5rVajVqu1qQRJKkO9Xqder7flWK3M\nyV8JrMrM5Y3XXwIyM7/W1OZx4JrM3Nl4/RTw3szcPeJYzslL0hHq9Jz8GmBRRJwdEUcDK4D7RrTZ\nSmOKpnHi9ZiRAS9J6r4JQz4zDwArgQeADcDqzNwYEXdExAcbzW4FPh0R64B/B27pVMGSpNZNOF3T\n1g9zukaSjlinp2skSTOUIS9JBTPkJalghrwkFcyQl6SCGfKSVDBDXpIKZshLUsEMeUkqmCEvSQUz\n5CWpYIa8JBXMkJekghnyklQwQ16SCmbIS1LBDHlJKpghL0kFM+QlqWCGvCQVzJCXpIIZ8pJUMENe\nkgpmyEtSwQx5SSqYIS9JBTPkJalghrwkFcyQl6SCGfKSVDBDXpIKZshLUsEMeUkqmCEvSQUz5CWp\nYC2FfEQsj4hNEbE5Im47TJuPRsSGiHgsIu5tb5mSpMmIzBy/QcQcYDNwNfAssAZYkZmbmtosAn4E\nvD8z90bEaZm5e4xj5USfJ0kaLiLIzJjM97Yykr8C2JKZWzNzP7AauH5Em08D387MvQBjBbwkqfta\nCfn5wPam1zsa7zW7ALgwIh6KiN9GxDXtKlCSNHnz2nicRcD7gLOA30TE0sGRvSSpN1oJ+Z1UwT1o\nQeO9ZjuAhzPzIPBMRGwGzgf6Rx5s1apVh57XajVqtdqRVSxJhavX69Tr9bYcq5UTr3OBJ6lOvO4C\nfgfclJkbm9pc03jvryLiNKpwvzQzXx5xLE+8StIR6uiJ18w8AKwEHgA2AKszc2NE3BERH2y0+QXw\nYkRsAP4XuHVkwEuSum/CkXxbP8yRvCQdsU4voZQkzVCGvCQVzJCXpIIZ8pJUMENekgpmyEtSwQx5\nSSqYIS9JBTPkJalghrwkFcyQl6SCGfKSVDBDXpIKZshLUsEMeUkqmCEvSQUz5CWpYIa8JBXMkJek\nghnyklQwQ16SCmbIS1LBDHlJKpghL0kFM+QlqWCGvCQVzJCXpIIZ8pJUMENekgpmyEtSwQx5SSqY\nIS9JBTPkJalghrwkFcyQl6SCGfKSVLCWQj4ilkfEpojYHBG3jdPuIxFxMCL62leiJGmyJgz5iJgD\n/BNwDbAEuCkiLhqj3fHA54GH212kJGlyWhnJXwFsycytmbkfWA1cP0a7rwBfBd5sY32SpCloJeTn\nA9ubXu9ovHdIRFwGLMjM+9tYmyRpiuZN9QAREcCdwC3Nb0/1uJKkqWsl5HcCZzW9XtB4b9AJVHP1\n9Ubgnwn8NCI+nJkDIw+2atWqQ89rtRq1Wu3Iq5akgtXrder1eluOFZk5foOIucCTwNXALuB3wE2Z\nufEw7X8F/H1mrh3jaznR50mShosIMnNSMyQTzsln5gFgJfAAsAFYnZkbI+KOiPjgWN+C0zWSNC1M\nOJJv64c5kpekI9bRkbwkaeYy5CWpYIa8JBXMkJekghnyklQwQ16SCmbIS1LBDHlJKpghL0kFM+Ql\nqWCGvCQVzJCXpIIZ8pJUMENekgpmyEtSwQx5SSqYIS9JBTPkJalghrwkFcyQl6SCGfKSVDBDXpIK\nZshLUsEMeUkqmCEvSQUz5CWpYF0P+e98Bx5+GF57rdufLEmzz7xuf+C6dXDPPfDEE3DuudDXN/S4\n9FI48cRuVyRJ5YrM7N6HReTg5+3bBxs2wMAA9PfD2rWwfj3Mnw+XXz4U/JddBqec0rUSJWnaiQgy\nMyb1vb0K+bG89RY8+eRQ6A8MVP+eeurwEX9fH5xxRtfKlqSeKibkx3LwIDz11PAR/8AAHHtsNcpv\nDv6FCyEm1Q2SNH0VHfJjyYRt24ZCv7+/ehw8OHrEf+65Br+kmW3WhfxYMmHXrmqU3zzq37t39Ij/\nggtg7tyOlCFJbWfIj+OFF4bm9gfD//nn4ZJLhp/gXbwYjjqqq6VJUksM+SO0Z0+1lLN5umfbNliy\nZGhFT18fXHxxNfcvSb1kyLfBK69Uwd884t+ypZra6esbGvUvWwbHHdfraiXNJh0P+YhYDnyD6grZ\nuzPzayO+/gXgb4D9wAvAX2fm9jGOM21DfixvvAGPPz58jn/DBjjrrOHBf9llcNJJva5WUqk6GvIR\nMQfYDFwNPAusAVZk5qamNlcBj2TmGxHxWaCWmSvGONaMCvmx7N9fBf3atUNTPevXV+v2R67sOe20\nXlcrqQSdDvkrgdsz89rG6y8BOXI039T+UuCuzPzTMb4240N+LAcOwObNo9fyn3TS6OB/5ztd0inp\nyEwl5Fu5d818oHnqZQdwxTjtPwXcP5liZqq5c6vVOYsXw803V+8dPAh/+MNQ4N91V/ULYN680cF/\n9tkGv6TOaOsNyiLi48DlwFWHa7Nq1apDz2u1GrVarZ0lTBtz5sCiRdXjxhur9zJhx46hEf8998DK\nldXc/8jgX7SoOoak2ader1Ov19tyrFana1Zl5vLG6zGnayLiA8A3gfdl5ouHOVaR0zVTtWvX0Bz/\n4C+Al1+u7srZHPwXXlj9JSBpdun0nPxc4EmqE6+7gN8BN2XmxqY2lwE/Bq7JzKfGOZYh36KXXho6\nsTt4Fe/OndVFXM3Bv2QJHH10r6uV1EndWkL5TYaWUH41Iu4A1mTmzyLil8BSql8CAWzNzBvGOI4h\nPwV791Zr+QdDf2CgmvdfvHh48F9yCbztbb2uVlK7eDHULPbaa9USzuZVPZs2wXnnjd6Q5YQTel2t\npMkw5DXMm28ObcgyOMf/+OOwYMHw4HdDFmlmMOQ1obfeqkb4zffrWbeuumCr+erdvj44/fReVyup\nmSGvSTl4sLqIq3lVz9q11b15Ri7pnD/ftfxSrxjyaptMeOaZ4aHf3199bWTwn3OOwS91gyGvjsqs\nlm82j/gHBuDVV6t5/eapnvPP9yIuqd0MefXE888PjfQH/33hhdEXcS1e7EVc0lQY8po2Xn55+Fr+\n/n7Yvh2WLh0e/EuXwjHH9LpaaWYw5DWtvfIKPPro8Kt3t2ypbtPQPNWzbBm8/e29rlaafgx5zTiv\nvw6PPTZ8xL9xY3Uyd+RFXG7IotnOkFcR9u2DJ54YfvXu+vXVPfhHXsTlhiyaTQx5FevAgeoiruaV\nPevWwcknDw/+yy+HM8/sdbVSZxjymlUGN2RpnuMfGKhO5I5cy79woWv5NfMZ8pr1MmHbttFX7+7b\nNzr4zz3XtfyaWQx56TB27Ro+2h8YgD17qnn95qmeCy6otnGUpiNDXjoCu3ePHvHv2jV6Q5b3vMcN\nWTQ9GPLSFP3xj9UJ3eaVPU8/XQV9c/BffLEbsqj7DHmpA159tVrC2Tzif/LJapP15qmeZcvg+ON7\nXa1KZshLXfLGG9UGLM0j/g0b4KyzRq/lP/nkXlerUhjyUg/t319drdt8cvfRR6vNV0au7HnHO3pd\nrWYiQ16aZg4cqO7P0zziHxio9tltnurp66uu6HUtv8ZjyEszQGZ1Mrf5fj0DA9XSzZEj/rPPNvg1\nxJCXZqjBDVmaR/z9/dUN3Abn9t2QRYa8VJjnnhu9E9fu3UMbsgxO9Vx0kRuyzAaGvDQLvPRSFfbN\nc/w7dlRr95unepYscUOW0hjy0iy1d2+1kqd5Lf9TT1Uj/Obgv+QSN2SZyQx5SYe89lp1EVfziH/j\nxurGbCM3ZDnxxF5Xq1YY8pLGtW9fddFW84h//XpYsGD0RVynnNLrajWSIS/piL311vANWQbn+089\ndfSSzjPO6HW1s5shL6ktDh6s5vSbR/z9/XDsscNX9fT1VX8FuJa/Owx5SR2TCVu3jl7Lnzl8Hf/g\nhiwGf/sZ8pK6KnNoQ5bm8N+7d3TwuyHL1BnykqaFF14Yvpa/v7+6sGvZstEbshx1VK+rnTkMeUnT\n1p49QxuyDJ7g3batumirOfiXLq3m/jWaIS9pRnnllSr4164dOrm7ZUs1tdMc/MuWwXHH9bra3jPk\nJc14gxuyNN+vZ8MGePe7R1/ENds2ZOl4yEfEcuAbwBzg7sz82oivHw38ALgc2A18LDO3jXEcQ15S\ny/bvr4K+eY5//Xo488zRa/lPO63X1XZOR0M+IuYAm4GrgWeBNcCKzNzU1OZzwMWZ+bcR8THgzzNz\nxRjHMuQb6vU6tVqt12VMC/bFEPtiyOH64sAB2Lx5+MqetWurWzSMDP53vav7dXfCVEK+lbtTXwFs\nycytmbkfWA1cP6LN9cC/Np7/J9UvBI2jXq/3uoRpw74YYl8MOVxfzJ0LixfDzTfDnXfCr35V3aHz\nwQer915/He66q7op25lnwnXXwZe/DD/5SbXef7aNM1u5E/V8YHvT6x1UwT9mm8w8EBF7IuKUzHyp\nPWVK0uHNmQOLFlWPG2+s3susbsXc31897rkHVq6EN98cPeI/77xyN2Tp1HYDXvMmqaciYOHC6nHD\nDUPv79o1NMf/ox/BF79YLfMc3JBl8HHhhWVsyNLKnPyVwKrMXN54/SUgm0++RsT9jTaPRMRcYFdm\nnj7GsWbZH0qS1B6TnZNv5ffUGmBRRJwN7AJWADeNaPPfwC3AI8CNwIPtLFKSNDkThnxjjn0l8ABD\nSyg3RsQdwJrM/BlwN/BvEbEFeJHqF4Ekqce6ejGUJKm7OnI+OSKWR8SmiNgcEbeN8fWjI2J1RGyJ\niP+LiLM6Ucd00EJffCEiNkTEuoj4ZUQs7EWd3TBRXzS1+0hEHIyIvm7W102t9EVEfLTxs/FYRNzb\n7Rq7pYX/Iwsj4sGIGGj8P7m2F3V2WkTcHRHPRcT6cdp8q5Gb6yLi0pYOnJltfVD94vg9cDZwFLAO\nuGhEm88B32k8/xiwut11TIdHi31xFXBs4/lnZ3NfNNodD/wa+C3Q1+u6e/hzsQjoB05svD6t13X3\nsC++B3ym8Xwx8HSv6+5QX/wJcCmw/jBfvxb4n8bz9wIPt3LcTozkvXhqyIR9kZm/zsw3Gi8fprrm\noESt/FwAfAX4KvBmN4vrslb64tPAtzNzL0Bm7u5yjd3SSl8cBAa3HD8Z2NnF+romMx8CXh6nyfVU\nt48hMx8BToqICTdm7ETIj3Xx1MjgGnbxFLAnIkrcPriVvmj2KeD+jlbUOxP2RURcBizIzFL7YFAr\nPxcXABdGxEMR8duIuKZr1XVXK31xB/CXEbEd+Bnwd12qbboZ2Vc7aWFQOF2W+s/6pZUR8XGqG7xd\n1etaeiEiAriTainuobd7VM50MI9qyuZ9wFnAbyJi6eDIfpa5Cfh+Zn69cd3OvcCSHtc0Y3RiJL+T\n6ody0AJG/3m1A1gI0Lh46sQs8xYIrfQFEfEB4B+ADzX+ZC3RRH1xAtV/3HpEPA1cCfy00JOvrf4f\nuS8zD2bmM1Q3CTy/O+V1VSt98SngPwAy82Hg2Igo+J6Th7WTRm42jJknI3Ui5A9dPNW4BfEK4L4R\nbQYvnoJxLp4qwIR90Zii+C7w4cx8sQc1dsu4fZGZezPz9Mw8NzPPoTo/8aHMHOhRvZ3Uyv+R/wLe\nD9AItPOBP3S1yu5opS+2Ah8AiIjFwDEFn6MIDv8X7H3AJ+DQnQj2ZOZzEx2w7dM16cVTh7TYF/8I\nHAf8uDFlsTUzbzj8UWemFvti2LdQ6HRNK32Rmb+IiD+LiA3AW8CtmTneSbkZqcWfi1uBf4mIL1Cd\nhL3l8EecuSLih0ANODUitgG3A0dT3UbmnzPz5xFxXUT8HngV+GRLx20sx5EkFajQm2tKksCQl6Si\nGfKSVDBDXpIKZshLUsEMeUkqmCEvSQUz5CWpYP8PFfRlnpBqyFUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe5676d86d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "k = np.linspace(0,1,100)\n",
    "plt.plot(k,-(w[0]/w[1])*k + theta/w[1] - w[2]/w[1])\n",
    "plt.xlim((0,1))\n",
    "plt.ylim((0,1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Let's use it for Iris!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 5.1  3.5]: -14.3926566688 -> 0\n",
      "[ 4.9  3. ]: -7.50863963457 -> 0\n",
      "[ 4.7  3.2]: -14.6948907479 -> 0\n",
      "[ 4.6  3.1]: -14.267939691 -> 0\n",
      "[ 5.   3.6]: -17.9857822255 -> 0\n",
      "[ 5.4  3.9]: -17.6835481464 -> 0\n",
      "[ 4.6  3.4]: -20.2980546114 -> 0\n",
      "[ 5.   3.4]: -13.9657056119 -> 0\n",
      "[ 4.4  2.9]: -13.4140375771 -> 0\n",
      "[ 4.9  3.1]: -9.51867794137 -> 0\n",
      "[ 5.4  3.7]: -13.6634715328 -> 0\n",
      "[ 4.8  3.4]: -17.1318801116 -> 0\n",
      "[ 4.8  3. ]: -9.09172688445 -> 0\n",
      "[ 4.3  3. ]: -17.0071631338 -> 0\n",
      "[ 5.8  4. ]: -13.3612374537 -> 0\n",
      "[ 5.7  4.4]: -22.9844779307 -> 0\n",
      "[ 5.4  3.9]: -17.6835481464 -> 0\n",
      "[ 5.1  3.5]: -14.3926566688 -> 0\n",
      "[ 5.7  3.8]: -10.92424809 -> 0\n",
      "[ 5.1  3.8]: -20.4227715892 -> 0\n",
      "[ 5.4  3.4]: -7.63335661239 -> 0\n",
      "[ 5.1  3.7]: -18.4127332824 -> 0\n",
      "[ 4.6  3.6]: -24.318131225 -> 0\n",
      "[ 5.1  3.3]: -10.3725800552 -> 0\n",
      "[ 4.8  3.4]: -17.1318801116 -> 0\n",
      "[ 5.  3.]: -5.9255523847 -> 0\n",
      "[ 5.   3.4]: -13.9657056119 -> 0\n",
      "[ 5.2  3.5]: -12.8095694189 -> 0\n",
      "[ 5.2  3.4]: -10.7995311121 -> 0\n",
      "[ 4.7  3.2]: -14.6948907479 -> 0\n",
      "[ 4.8  3.1]: -11.1017651912 -> 0\n",
      "[ 5.4  3.4]: -7.63335661239 -> 0\n",
      "[ 5.2  4.1]: -24.8697992597 -> 0\n",
      "[ 5.5  4.2]: -22.1305758169 -> 0\n",
      "[ 4.9  3.1]: -9.51867794137 -> 0\n",
      "[ 5.   3.2]: -9.94562899829 -> 0\n",
      "[ 5.5  3.5]: -8.06030766931 -> 0\n",
      "[ 4.9  3.1]: -9.51867794137 -> 0\n",
      "[ 4.4  3. ]: -15.4240758839 -> 0\n",
      "[ 5.1  3.4]: -12.382618362 -> 0\n",
      "[ 5.   3.5]: -15.9757439187 -> 0\n",
      "[ 4.5  2.3]: 0.229279513512 -> 0\n",
      "[ 4.4  3.2]: -19.4441524975 -> 0\n",
      "[ 5.   3.5]: -15.9757439187 -> 0\n",
      "[ 5.1  3.8]: -20.4227715892 -> 0\n",
      "[ 4.8  3. ]: -9.09172688445 -> 0\n",
      "[ 5.1  3.8]: -20.4227715892 -> 0\n",
      "[ 4.6  3.2]: -16.2779779978 -> 0\n",
      "[ 5.3  3.7]: -15.2465587827 -> 0\n",
      "[ 5.   3.3]: -11.9556673051 -> 0\n",
      "[ 7.   3.2]: 21.7161159992 -> 1\n",
      "[ 6.4  3.2]: 12.2175925 -> 1\n",
      "[ 6.9  3.1]: 22.1430670561 -> 1\n",
      "[ 5.5  2.3]: 16.0601520123 -> 1\n",
      "[ 6.5  2.8]: 21.840832977 -> 1\n",
      "[ 5.7  2.8]: 9.17613497802 -> 1\n",
      "[ 6.3  3.3]: 8.62446694329 -> 1\n",
      "[ 4.9  2.4]: 4.55159020621 -> 1\n",
      "[ 6.6  2.9]: 21.4138819201 -> 1\n",
      "[ 5.2  2.7]: 3.27073703545 -> 1\n",
      "[ 5.  2.]: 14.1748306833 -> 1\n",
      "[ 5.9  3. ]: 8.32223286418 -> 1\n",
      "[ 6.   2.2]: 25.9856265684 -> 1\n",
      "[ 6.1  2.9]: 13.4984456707 -> 1\n",
      "[ 5.6  2.9]: 5.58300942135 -> 1\n",
      "[ 6.7  3.1]: 18.9768925564 -> 1\n",
      "[ 5.6  3. ]: 3.57297111455 -> 1\n",
      "[ 5.8  2.7]: 12.7692605347 -> 1\n",
      "[ 6.2  2.2]: 29.1518010682 -> 1\n",
      "[ 5.6  2.5]: 13.6231626485 -> 1\n",
      "[ 5.9  3.2]: 4.30215625058 -> 1\n",
      "[ 6.1  2.8]: 15.5084839775 -> 1\n",
      "[ 6.3  2.5]: 24.7047733977 -> 1\n",
      "[ 6.1  2.8]: 15.5084839775 -> 1\n",
      "[ 6.4  2.9]: 18.2477074204 -> 1\n",
      "[ 6.6  3. ]: 19.4038436133 -> 1\n",
      "[ 6.8  2.8]: 26.5900947266 -> 1\n",
      "[ 6.7  3. ]: 20.9869308632 -> 1\n",
      "[ 6.   2.9]: 11.9153584209 -> 1\n",
      "[ 5.7  2.6]: 13.1962115916 -> 1\n",
      "[ 5.5  2.4]: 14.0501137055 -> 1\n",
      "[ 5.5  2.4]: 14.0501137055 -> 1\n",
      "[ 5.8  2.7]: 12.7692605347 -> 1\n",
      "[ 6.   2.7]: 15.9354350344 -> 1\n",
      "[ 5.4  3. ]: 0.406796614803 -> 1\n",
      "[ 6.   3.4]: 1.86516688686 -> 1\n",
      "[ 6.7  3.1]: 18.9768925564 -> 1\n",
      "[ 6.3  2.3]: 28.7248500113 -> 1\n",
      "[ 5.6  3. ]: 3.57297111455 -> 1\n",
      "[ 5.5  2.5]: 12.0400753987 -> 1\n",
      "[ 5.5  2.6]: 10.0300370919 -> 1\n",
      "[ 6.1  3. ]: 11.4884073639 -> 1\n",
      "[ 5.8  2.6]: 14.7792988415 -> 1\n",
      "[ 5.   2.3]: 8.14471576289 -> 1\n",
      "[ 5.6  2.7]: 9.60308603495 -> 1\n",
      "[ 5.7  3. ]: 5.15605836443 -> 1\n",
      "[ 5.7  2.9]: 7.16609667123 -> 1\n",
      "[ 6.2  2.9]: 15.0815329206 -> 1\n",
      "[ 5.1  2.5]: 5.70772639917 -> 1\n",
      "[ 5.7  2.8]: 9.17613497802 -> 1\n"
     ]
    }
   ],
   "source": [
    "import itertools\n",
    "training_data = X\n",
    "\n",
    "training_data = [np.append(x,[1]) for x in training_data]\n",
    "target = Y\n",
    "w = random.rand(3) # Initializing weights randomly.\n",
    "errors = []\n",
    "eta = 0.2 # Learning rate\n",
    "n = 100000 # Number of iterations\n",
    "theta = 0.25 # Threshold\n",
    "\n",
    "data = itertools.cycle(zip(training_data,target))\n",
    "\n",
    "for i in range(n):\n",
    "    x, expected = next(data)    \n",
    "    result = dot(w, x)\n",
    "    error = expected - unit_step(result,theta)\n",
    "    errors.append(error)\n",
    "    w += eta * error * x\n",
    "\n",
    "for x, _ in zip(training_data,target):\n",
    "    result = dot(x, w)\n",
    "    print(\"{}: {} -> {}\".format(x[:2], result, unit_step(result,theta)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7fe54d51bd90>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEPCAYAAABY9lNGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLrxlzgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe54d53a710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(X[:, 0], X[:, 1], c=Y, cmap=plt.cm.Paired)\n",
    "plt.plot(X[:,0],-(w[0]/w[1])*X[:,0] + theta/w[1] - w[2]/w[1])\n",
    "plt.xlabel('Sepal length')\n",
    "plt.ylabel('Sepal width')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe54cd1b190>]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEACAYAAACgS0HpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe54d4b14d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from pylab import plot, ylim\n",
    "ylim([-1,1])\n",
    "plot(errors)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Perceptron calculation\n",
    "plant_class = ((X[:,0] * w[0] + X[:,1] * w[1] + w[2]) > theta) + 0\n",
    "\n",
    "# Compare\n",
    "correctly_classified = np.all(plant_class == Y)\n",
    "correctly_classified\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7875905869660299"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "-(w[0]/w[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-1.2451884916755898"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    " theta/w[1] - w[2]/w[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "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": 1
}