Neural Network by JAKK-coder.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generating Random Data Set For model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.vstack([(np.random.rand(10, 2)*5),(np.random.rand(10, 2)*10)])\n",
    "Y = np.hstack([[0]*10 , [1]*10])\n",
    "dataset = pd.DataFrame(X,columns={\"X1\",\"X2\"})\n",
    "dataset[\"Y\"] =Y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random DataSet Generated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>X1</th>\n",
       "      <th>X2</th>\n",
       "      <th>Y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3.596683</td>\n",
       "      <td>2.528780</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.375634</td>\n",
       "      <td>1.299973</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.569377</td>\n",
       "      <td>1.768548</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.374849</td>\n",
       "      <td>3.473490</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.271004</td>\n",
       "      <td>4.970541</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         X1        X2  Y\n",
       "0  3.596683  2.528780  0\n",
       "1  1.375634  1.299973  0\n",
       "2  2.569377  1.768548  0\n",
       "3  2.374849  3.473490  0\n",
       "4  4.271004  4.970541  0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ploting dataset values "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1bf4fe78d68>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.plot(dataset,label='Inline label')\n",
    "plt.legend([\"X2\",\"X1\",\"Y\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Converting Target variable to vector of number of classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "Z = np.zeros((20, 2))\n",
    "for i in range(20):\n",
    "    Z[i, Y[i]] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20, 2)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Initiating Random Weights and Bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "Wi_1 = np.random.randn(3, 2)\n",
    "Bi_1 = np.random.randn(3)\n",
    "Wi_2 = np.random.randn(2, 3)\n",
    "Bi_2 = np.random.randn(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-1.56462436, -0.05221165],\n",
       "       [ 1.58601077,  0.13369813],\n",
       "       [-1.13341516, -0.38220891]])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Wi_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.41607498, 1.23941914])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Bi_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.19812185, -0.06233041,  1.22597364],\n",
       "       [-0.00875469, -1.52851499, -2.30988988]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Wi_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1.91687084, -0.26137672])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Bi_2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Defining the Forward Propagation function "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def forward_prop(X, Wi_1, Bi_1, Wi_2, Bi_2):\n",
    "    M = 1 / (1 + np.exp(-(X.dot(Wi_1.T) + Bi_1)))\n",
    "    A = M.dot(Wi_2.T) + Bi_2\n",
    "    expA = np.exp(A)\n",
    "    Y = expA / expA.sum(axis=1, keepdims=True)\n",
    "    return Y, M"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "M = 1 / (1 + np.exp(-(X.dot(Wi_1.T) + Bi_1)))\n",
    "A = M.dot(Wi_2.T) + Bi_2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Backpropagation by Differentiating or Gradient Descent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Returns gradient for Weight_2\n",
    "def diff_Wi_2(H, Z, Y):\n",
    "    return H.T.dot(Z - Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Returns gradient for Weight_2\n",
    "def diff_Wi_1(X, H, Z, output, Wi_2):\n",
    "    dZ = (Z - output).dot(Wi_2) * H * (1 - H)\n",
    "    return X.T.dot(dZ)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Returns derivative for both bias\n",
    "def diff_B2(Z, Y):\n",
    "    return (Z - Y).sum(axis=0)\n",
    "\n",
    "def diff_B1(Z, Y, Wi_2, H):\n",
    "    return ((Z - Y).dot(Wi_2) * H * (1 - H)).sum(axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate = 1e-3\n",
    "for epoch in range(5000):\n",
    "    output, hidden = forward_prop(X, Wi_1, Bi_1, Wi_2, Bi_2)\n",
    "    Wi_2 += learning_rate * diff_Wi_2(hidden, Z, output).T\n",
    "    Bi_2 += learning_rate * diff_B2(Z, output)\n",
    "    Wi_1 += learning_rate * diff_Wi_1(X, hidden, Z, output, Wi_2).T\n",
    "    Bi_1 += learning_rate * diff_B1(Z, output, Wi_2, hidden)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generating Prediction at New Datapoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " prob of class 0>>>>>>  0.7991464774581799 \n",
      " prob of class 1 >>>>> 0.20085352254182\n"
     ]
    }
   ],
   "source": [
    "X_test =np.array([1,1])\n",
    "\n",
    "hidden_output = 1 / (1 + np.exp(-X_test.dot(Wi_1.T) - Bi_1))\n",
    "Outer_layer_output = hidden_output.dot(Wi_2.T) + Bi_2\n",
    "expA = np.exp(Outer_layer_output)\n",
    "Y = expA / expA.sum()\n",
    "print(\" prob of class 0>>>>>>  {} \\n prob of class 1 >>>>> {}\".format(Y[0],Y[1]))"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}