q1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "q1.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 74,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "68u2App6SrVb",
        "outputId": "9c12874d-a451-4261-f73d-c97143b876c6"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import math\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "%matplotlib inline\n",
        "def make_spiral_xy(n,i):\n",
        "      φ = i/16 * math.pi\n",
        "      r = 6.5 * ((104 - i)/104)\n",
        "      x = (r * math.cos(φ) * n)/13 + 0.5\n",
        "      y = (r * math.sin(φ) * n)/13 + 0.5\n",
        "      plt.scatter(x,y)\n",
        "      return [x,y]\n",
        "\n",
        "X = [make_spiral_xy(1,i) for i in range(100)]\n",
        "Y = [make_spiral_xy(-1,i) for i in range(100)]\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class NeuralNetworkTanh(object):\n",
        "    def __init__(self):\n",
        "        inputLayerNeurons = 2\n",
        "        hiddenLayerNeurons = 10\n",
        "        outLayerNeurons = 1\n",
        "        \n",
        "        self.learning_rate = 0.2\n",
        "        self.W_HI = np.random.randn(inputLayerNeurons, hiddenLayerNeurons)\n",
        "        self.W_OH = np.random.randn(hiddenLayerNeurons, outLayerNeurons)\n",
        "        \n",
        "    def tanh(self, x, der=False):\n",
        "        if der == True:\n",
        "            return 1-x**2\n",
        "        else:\n",
        "            return (np.exp(2*x) - 1) / (np.exp(2*x) + 1)\n",
        "        \n",
        "    def feedForward(self, X):\n",
        "        hidden_input = np.dot(X, self.W_HI)\n",
        "        self.hidden_output = self.tanh(hidden_input)\n",
        "        \n",
        "        output_input = np.dot(self.hidden_output, self.W_OH)\n",
        "        pred = self.tanh(output_input)\n",
        "        return pred\n",
        "        \n",
        "    def backPropagation(self, X, Y, pred):\n",
        "        output_error = Y - pred\n",
        "        output_delta = self.learning_rate * output_error * self.tanh(pred, der=True)\n",
        "        \n",
        "        hidden_error = output_delta.dot(self.W_OH.T)\n",
        "        hidden_delta = self.learning_rate * hidden_error * self.tanh(self.hidden_output, der=True)\n",
        "        \n",
        "        self.W_HI += X.T.dot(hidden_delta)\n",
        "        self.W_OH += self.hidden_output.T.dot(output_delta)\n",
        "        \n",
        "    def train(self, X, Y):\n",
        "        output = self.feedForward(X)\n",
        "        self.backPropagation(X,Y,output)\n"
      ],
      "metadata": {
        "id": "0I8Y889fcbrk"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "NNT = NeuralNetworkTanh()\n",
        "err = []\n",
        "for i in range(10000):\n",
        "    #NN.train(X,Y)\n",
        "     err.append(np.mean(np.square(Y - NNT.feedForward(X))))\n",
        "\n",
        "plt.plot(err)\n",
        "print(NNT.feedForward([0,0]))\n",
        "print(NNT.feedForward([1,1]))\n",
        "print(NNT.feedForward([1,0]))\n",
        "print(NNT.feedForward([0,1]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 334
        },
        "id": "2woFYHlvfcoX",
        "outputId": "ef7cec5d-8dd0-45fd-db35-dd63ede9a1d3"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.]\n",
            "[-0.09199658]\n",
            "[-0.81694742]\n",
            "[0.64821233]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class NeuralNetworkRelu(object):\n",
        "    def __init__(self):\n",
        "        inputLayerNeurons = 2\n",
        "        hiddenLayerNeurons = 10\n",
        "        outLayerNeurons = 1\n",
        "        \n",
        "        self.learning_rate = 0.2\n",
        "        self.W_HI = np.random.randn(inputLayerNeurons, hiddenLayerNeurons)\n",
        "        self.W_OH = np.random.randn(hiddenLayerNeurons, outLayerNeurons)\n",
        "        \n",
        "    def relu(self, x, der=False):\n",
        "        if der == True:\n",
        "            return 1. * (x > 0)\n",
        "        else:\n",
        "            return x * (x > 0)\n",
        "        \n",
        "    def feedForward(self, X):\n",
        "        hidden_input = np.dot(X, self.W_HI)\n",
        "        self.hidden_output = self.relu(hidden_input)\n",
        "        \n",
        "        output_input = np.dot(self.hidden_output, self.W_OH)\n",
        "        pred = self.relu(output_input)\n",
        "        return pred\n",
        "        \n",
        "    def backPropagation(self, X, Y, pred):\n",
        "        output_error = Y - pred\n",
        "        output_delta = self.learning_rate * output_error * self.relu(pred, der=True)\n",
        "        \n",
        "        hidden_error = output_delta.dot(self.W_OH.T)\n",
        "        hidden_delta = self.learning_rate * hidden_error * self.relu(self.hidden_output, der=True)\n",
        "        \n",
        "        self.W_HI += X.T.dot(hidden_delta)\n",
        "        self.W_OH += self.hidden_output.T.dot(output_delta)\n",
        "        \n",
        "    def train(self, X, Y):\n",
        "        output = self.feedForward(X)\n",
        "        self.backPropagation(X,Y,output)\n",
        "\n"
      ],
      "metadata": {
        "id": "rOjFgLiNfVdq"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "NNR = NeuralNetworkTanh()\n",
        "err = []\n",
        "for i in range(10000):\n",
        "    #NN.train(X,Y)\n",
        "     err.append(np.mean(np.square(Y - NNR.feedForward(X))))\n",
        "\n",
        "plt.plot(err)\n",
        "print(NNR.feedForward([0,0]))\n",
        "print(NNR.feedForward([1,1]))\n",
        "print(NNR.feedForward([1,0]))\n",
        "print(NNR.feedForward([0,1]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 334
        },
        "id": "uFYiglhXfvMR",
        "outputId": "7e9d82c2-4473-4d1c-b4ad-147733267a20"
      },
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0.]\n",
            "[-0.26492266]\n",
            "[0.16116465]\n",
            "[-0.93686565]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}