RDPTA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Name : Omkar Thawakar\n",
    "Reg No : 2015BCS003 , Roll No : A-08\n",
    "#### Aim : Implement R- Category Descrete Perceptron clasifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from matplotlib.colors import ListedColormap\n",
    "\n",
    "class RDPTA:\n",
    "    def __init__(self,neurons,eta=1):\n",
    "        self.neurons=neurons\n",
    "        self.eta=eta\n",
    "        \n",
    "    def init_weights(self,X,weights):\n",
    "    \tself.weights = np.array(weights)\n",
    "    \t'''if weights:\n",
    "    \t    self.weights = np.array(weights)\n",
    "    \telse:\n",
    "    \t    self.weights =  np.random.random((1+np.array(X).shape[1],self.neurons))\n",
    "    \t'''\n",
    "    def aug_inputs(self,X):\n",
    "        for x in X:\n",
    "            x.append(1)\n",
    "        return X\n",
    "\n",
    "    \n",
    "    def train(self,X,outputs,epochs):\n",
    "        self.weights =self.weights\n",
    "        print(\"Initial Weights : \",self.weights)\n",
    "        X = self.aug_inputs(X)\n",
    "        print(\"Augmented input pattern : \",X)\n",
    "        self.error_ = []\n",
    "        epoch=1\n",
    "        p = 0\n",
    "        for i in range(epochs):\n",
    "            print(\"Epoch : \",epoch)\n",
    "            err=0\n",
    "            for x,d in zip(X,outputs):\n",
    "                out = self.predict(x)\n",
    "                #print(\"D : \",d)\n",
    "                print(\"For input pattern : \",x)\n",
    "                print(\"Output of network : \",out)\n",
    "                err += 0.5*(d-out)**2\n",
    "                delta_=[]\n",
    "                for d_,o_ in zip(d,out):\n",
    "                    delta_.append(0.5*self.eta*(d_-o_))\n",
    "                #print(\"Delta : \",delta_)\n",
    "                del_w = []\n",
    "                for del_ in delta_:\n",
    "                    del_w.append(np.multiply(del_,x))\n",
    "                #print(\"Delta_W is : \",del_w)\n",
    "                for i in range(len(self.weights)):\n",
    "                    self.weights[i] = self.weights[i]+del_w[i]    \n",
    "                #self.weights = self.weights+ 0.5*self.eta*(d-out)*x\n",
    "                print(\"Weights : \",self.weights)\n",
    "                p+=1\n",
    "                self.error_.append(err)\n",
    "            print(\"Error : \",err)\n",
    "        print(\"Training Steps Required are {} .\".format(p))\n",
    "        return self\n",
    "                                                                           \n",
    "    def get_weights(self):\n",
    "        return self.weights\n",
    "                                                                           \n",
    "    def net_input(self,X):\n",
    "        return np.dot(self.weights,X)\n",
    "    \n",
    "    def predict(self,x):\n",
    "        print(\"Net : \",self.net_input(np.array(x)))\n",
    "        return np.where(self.net_input(np.array(x))>0,1,-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial Weights of Neuron are :  [[0 0 0]\n",
      " [0 0 0]\n",
      " [0 0 0]]\n",
      "Initial Weights :  [[0 0 0]\n",
      " [0 0 0]\n",
      " [0 0 0]]\n",
      "Augmented input pattern :  [[0.8, 0.5, 0, -1], [0, 0.2, 0.3, -1], [0.9, 0.7, 0.3, -1], [0.2, 0.1, 1.3, -1], [1, 0.8, 0.5, -1], [0.2, 0.7, 0.8, -1]]\n",
      "Epoch :  1\n"
     ]
    },
    {
     "ename": "ValueError",
     "evalue": "shapes (3,3) and (4,) not aligned: 3 (dim 1) != 4 (dim 0)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-aaabf0244544>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0mneurons\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minit_weights\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mweights\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Initial Weights of Neuron are : \"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mneurons\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweights\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0mneurons\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0md\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m4\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-1-0dcfe3c37c78>\u001b[0m in \u001b[0;36mtrain\u001b[0;34m(self, X, outputs, epochs)\u001b[0m\n\u001b[1;32m     34\u001b[0m             \u001b[0merr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0md\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 36\u001b[0;31m                 \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     37\u001b[0m                 \u001b[0;31m#print(\"D : \",d)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m                 \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"For input pattern : \"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-1-0dcfe3c37c78>\u001b[0m in \u001b[0;36mpredict\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     64\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     65\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 66\u001b[0;31m         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Net : \"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnet_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     67\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwhere\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnet_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m>\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-1-0dcfe3c37c78>\u001b[0m in \u001b[0;36mnet_input\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m     61\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mnet_input\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 63\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweights\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     64\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     65\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: shapes (3,3) and (4,) not aligned: 3 (dim 1) != 4 (dim 0)"
     ]
    }
   ],
   "source": [
    "X = [[0.8,0.5,0],[0,0.2,0.3],[0.9,0.7,0.3],[0.2,0.1,1.3],[1,0.8,0.5],[0.2,0.7,0.8]]\n",
    "d = [[-1],[1],[-1],[1],[-1],[1]]\n",
    "weights = [[0,0,0],[0,0,0],[0,0,0]]\n",
    "\n",
    "neurons = RDPTA(1)\n",
    "neurons.init_weights(X,weights)\n",
    "print(\"Initial Weights of Neuron are : \",neurons.weights)\n",
    "neurons.train(X,d,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Final Weights of Neuron are : \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[ 5,  3,  5],\n",
       "       [ 0, -1,  2],\n",
       "       [-9,  1,  0]])"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Final Weights of Neuron are : \")\n",
    "neurons.weights"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualization of Decision Boundaries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd4bae56a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i,j in zip(X,d):\n",
    "    plt.scatter(i[0],i[1],c='k' ,marker='x' if j==[1,-1,-1]  else '*' if j==[-1,1,-1] else 'o')\n",
    "\n",
    "    \n",
    "x = np.arange(-10,15)\n",
    "a1,b1,c1 = neurons.weights[0]\n",
    "y1 = (c1-a1*x)/b1\n",
    "plt.plot(x,y1,c='g',label='5x+3y-5=0')\n",
    "\n",
    "a2,b2,c2 = neurons.weights[1]\n",
    "y2 = (c2-a2*x)/b2\n",
    "plt.plot(x,y2,c='b',label='x-y-2=0')\n",
    "\n",
    "a3,b3,c3 = neurons.weights[2]\n",
    "y3 = (c3-a3*x)/b3\n",
    "plt.plot(x,y3,c='r',label='-9x+y-0=0')\n",
    "\n",
    "#fill Indecision regions\n",
    "y4 = np.maximum(y1, y3)\n",
    "plt.fill_between(x,y4,150,color='grey',alpha=0.5,label='Indecision region')\n",
    "y5 = np.minimum(y1, y2)\n",
    "plt.fill_between(x,y5,y2,color='grey',alpha=0.5)\n",
    "y6 = np.minimum(y2, y3)\n",
    "plt.fill_between(x,y6,y2,color='grey',alpha=0.5)\n",
    "y7 = np.minimum(y1,y3)\n",
    "y8 = np.minimum(1.5,y2)\n",
    "plt.fill_between(x,y2,y7,color='grey',alpha=0.5)\n",
    "\n",
    "\n",
    "\n",
    "plt.ylim([-10,10])\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "shapes (3,3) and (4,) not aligned: 3 (dim 1) != 4 (dim 0)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-5-b868e3489955>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     92\u001b[0m         \u001b[0ms\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minit_params\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     93\u001b[0m         \u001b[0ms\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_weights\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mweights\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 94\u001b[0;31m         \u001b[0ms\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     95\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-5-b868e3489955>\u001b[0m in \u001b[0;36mtrain\u001b[0;34m(self, interactive)\u001b[0m\n\u001b[1;32m     53\u001b[0m                                         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'%20s : %s'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m'W (before adjust)'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mW\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     54\u001b[0m                                         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'%20s : %s'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m'Input'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maug_x\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 55\u001b[0;31m                                 \u001b[0mnet\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mW\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mT\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maug_x\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     56\u001b[0m                                 \u001b[0;32mif\u001b[0m \u001b[0mnet\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     57\u001b[0m                                         \u001b[0;31m# if self.y_input[step] == -1:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: shapes (3,3) and (4,) not aligned: 3 (dim 1) != 4 (dim 0)"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "class DiscretePerceptron:\n",
    "\n",
    "\tdef __init__(self, x_input, y_input, activation=None):\n",
    "\t\t'''\n",
    "\t\tx_input = data points\n",
    "\t\ty_input = target labels\n",
    "\t\t'''\n",
    "\t\tif len(x_input) != len(y_input):\n",
    "\t\t\traise ValueError('X and Y must have length.')\n",
    "\n",
    "\t\tself.n_dims = x_input.shape[-1]\n",
    "\n",
    "\t\tself.aug_x = self.augment_input(x_input)\n",
    "\n",
    "\t\tself.y_input = y_input\n",
    "\t\tif activation is None:\n",
    "\t\t\tself.__activation = self.__threshold\n",
    "\n",
    "\tdef init_params(self):\n",
    "\t\tself.P = len(self.aug_x)\n",
    "\t\tself.error = 0\n",
    "\t\tself.p = 1\n",
    "\t\tself.k = 1\n",
    "\t\tself.c = 1 # learning rate\n",
    "\t\tself.W = np.random.random((self.aug_x.shape[-1]))\n",
    "\n",
    "\tdef augment_input(self, x_input):\n",
    "\t\t_input = []\n",
    "\t\tfor i in x_input:\n",
    "\t\t\t_input.append(np.concatenate((i, [1])))\n",
    "\t\treturn np.array(_input)\n",
    "\n",
    "\tdef set_weights(self, W):\n",
    "\t\tself.W = np.array(W)\n",
    "\n",
    "\tdef __threshold(self, net):\n",
    "\t\tif net >= 0:\n",
    "\t\t\treturn 1\n",
    "\t\telse:\n",
    "\t\t\treturn -1\n",
    "\n",
    "\tdef train(self, interactive=False):\n",
    "\t\tstop = False\n",
    "\t\twhile not stop:\n",
    "\t\t\tif interactive:\n",
    "\t\t\t\tprint('Training Cycle #{}:'.format(self.k))\n",
    "\t\t\tself.E = 0\n",
    "\t\t\tfor step in range(self.P):\n",
    "\t\t\t\tif interactive:\n",
    "\t\t\t\t\tprint('Step #{}:'.format(self.p))\n",
    "\t\t\t\t\tprint('%20s : %s' % ('W (before adjust)', (self.W)))\n",
    "\t\t\t\t\tprint('%20s : %s' % ('Input', (self.aug_x[step])))\n",
    "\t\t\t\tnet = np.dot(self.W.T, self.aug_x[step])\n",
    "\t\t\t\tif net == 0:\n",
    "\t\t\t\t\t# if self.y_input[step] == -1:\n",
    "\t\t\t\t\t# \toutput = 1\n",
    "\t\t\t\t\t# elif self.y_input[step] == 1:\n",
    "\t\t\t\t\t# \toutput = -1\n",
    "\t\t\t\t\toutput = 0\n",
    "\t\t\t\telse:\n",
    "\t\t\t\t\toutput = self.__activation(net)\n",
    "\n",
    "\t\t\t\t_delta = self.c * 0.5 * (self.y_input[step] - output) * self.aug_x[step]\n",
    "\t\t\t\tself.W = self.W + (_delta)\n",
    "\t\t\t\tself.E = self.E + 0.5 * (self.y_input[step] - output) ** 2\n",
    "\t\t\t\tif interactive:\n",
    "\t\t\t\t\tprint('%20s : %s' % ('Outputs (d, O)', (self.y_input[step], output)))\n",
    "\t\t\t\t\tprint('%20s : %s' % ('Error', self.E))\n",
    "\t\t\t\t\tprint('%20s : %s' % ('W (after adjust)', self.W))\n",
    "\t\t\t\tself.p += 1\n",
    "\n",
    "\t\t\tif self.E == 0:\n",
    "\t\t\t\tstop = True\n",
    "\t\t\tself.k += 1\n",
    "\t\t\tif interactive:\n",
    "\t\t\t\tinput()\n",
    "\n",
    "\t\tprint('Training steps required: {}'.format(self.p))\n",
    "\n",
    "if __name__=='__main__':\n",
    "\t# Testing for AND\n",
    "\n",
    "\tX = [[0.8,0.5,0],[0,0.2,0.3],[0.9,0.7,0.3],[0.2,0.1,1.3],[1,0.8,0.5],[0.2,0.7,0.8]]\n",
    "\td = [-1,1,-1,1,-1,1]\n",
    "\tweights = [[0,0,0],[0,0,0],[0,0,0]]\n",
    "\n",
    "\tx = np.array(X)\n",
    "\ty = np.array(d)\n",
    "\ts = DiscretePerceptron(x, y)\n",
    "\ts.init_params()\n",
    "\ts.set_weights(weights)\n",
    "\ts.train()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}