HW4_Part2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "HW4-Part2.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "xYBu4IT2hTNQ",
        "eSckmMrIhZoE"
      ],
      "authorship_tag": "ABX9TyODJX2zj2fmqtelkuc26wiQ",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/natashia-98/DataMining/blob/main/HW4_Part2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Implementing Linear Regression**"
      ],
      "metadata": {
        "id": "xYBu4IT2hTNQ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HSP0X7lzYdq9"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "learning_rate = 0.01\n",
        "epochs = 200"
      ],
      "metadata": {
        "id": "9nRQT3XnYiaC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Get training data"
      ],
      "metadata": {
        "id": "e-1esV7EZbVL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "n_samples = 70\n",
        "train_x = np.linspace(0, 20, n_samples)\n",
        "train_y = 3 * train_x + 5 * np.random.randn(n_samples)"
      ],
      "metadata": {
        "id": "LaPeAEfRYlAD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(train_x, train_y, 'o')\n",
        "plt.plot(train_x, 3 * train_x)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "WKnFHRQ1YvwJ",
        "outputId": "eb70e23d-0f00-4aad-e9f5-bc65396c203f"
      },
      "execution_count": null,
      "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"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "Model Implementation"
      ],
      "metadata": {
        "id": "sGXu1yLpZt3G"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X = tf.placeholder(tf.float32)\n",
        "Y = tf.placeholder(tf.float32)\n",
        "B1 = tf.Variable(np.random.randn())\n",
        "B0 = tf.Variable(np.random.randn())\n",
        "lrm = tf.add(tf.multiply(B1,X),B0)\n",
        "cost = tf.reduce_sum((lrm - Y) ** 2) / (2 * n_samples)\n",
        "optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)"
      ],
      "metadata": {
        "id": "wucTXM9bZaXY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "init = tf.global_variables_initializer()"
      ],
      "metadata": {
        "id": "zDfxtOBza97r"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with tf.Session() as tses:\n",
        "  tses.run(init)\n",
        "\n",
        "  for i in range(epochs):\n",
        "    for x,y in zip(train_x,train_y):\n",
        "      tses.run(optimizer, feed_dict = {X:x, Y:y})\n",
        "  weight1 = tses.run(B1)\n",
        "  weight2 = tses.run(B0)\n",
        "    "
      ],
      "metadata": {
        "id": "gWemQ63pdTPr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Test model outcome"
      ],
      "metadata": {
        "id": "4BT24F4jZy87"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(weight1, weight2)\n",
        "\n",
        "plt.plot(train_x,train_y, 'o')\n",
        "plt.plot(train_x, weight1*train_x + weight2)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "3D9lXme1fIGG",
        "outputId": "1303cdd8-7a7c-4340-e965-086d62cb3da9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3.171673 -2.307742\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": "markdown",
      "source": [
        "Our model produced a line which has weights which are close to our expected values, so our model works (weight2 represents bias)"
      ],
      "metadata": {
        "id": "8sBlhQoWZ26P"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Implementing Logistic Regression**"
      ],
      "metadata": {
        "id": "eSckmMrIhZoE"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**In this code, we will try to classify data from the iris dataset**"
      ],
      "metadata": {
        "id": "DVsu2AJGhxNC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "In order for this code to properly execute, first do the following if you have already executed the previous codes:\n",
        "Runtime -> Restart runtime"
      ],
      "metadata": {
        "id": "ZhwBAkKCXurz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import datasets\n",
        "from sklearn.datasets import load_iris\n",
        "import numpy\n"
      ],
      "metadata": {
        "id": "N9gJDL7XFS33"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "learning_rate = 0.1\n",
        "epochs = 200\n",
        "\n",
        "iris = load_iris(as_frame = True)\n",
        "df = iris['data'].join(iris['target'])\n",
        "df['Setosa'] = (df['target'] == 0)*1.0\n",
        "df['Versicolor'] = (df['target'] == 1)*1.0\n",
        "df['Virginica'] = (df['target'] == 2)*1.0\n",
        "train_data = df.sample(frac=0.8, random_state=0)\n",
        "test_data = df.drop(train_data.index)"
      ],
      "metadata": {
        "id": "h2wJW3brFTpk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Normalize data"
      ],
      "metadata": {
        "id": "zfvSUN8pFZlN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn import preprocessing\n",
        "norm_train=(train_data.iloc[:,0:4]-train_data.iloc[:,0:4].mean())/train_data.iloc[:,0:4].std()\n",
        "norm_test=(test_data.iloc[:,0:4]-test_data.iloc[:,0:4].mean())/test_data.iloc[:,0:4].std()\n",
        "train_y = train_data['target']\n",
        "test_y = test_data['target']\n",
        "train_x1 = norm_train.iloc[:,2:3]\n",
        "test_x1 = norm_test.iloc[:,2:3]\n"
      ],
      "metadata": {
        "id": "1MiJuulrFWVP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Visualize data"
      ],
      "metadata": {
        "id": "x_6o1hzJaQbD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(train_x1, train_y, 'o')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "UHU1utIlFeRq",
        "outputId": "eb372d1f-7b7c-49fb-e11e-c879320e82be"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD4CAYAAADiry33AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAYYklEQVR4nO3df3Bd5X3n8ffHF9nWUlPbsXBA2BgYLw2tYzvcsZ3CJM5usV2nix0Cg72lgYbUTRvS6WaGGShMmKFmYOvZLOyGDvFQN9C0JtMEXHc31HGasrTp2uWKH1GAOBhDYguK1Tj8Cl6wxXf/uMfmSrpX90j3Sld69HnNaHTPc55zzlfnSJ97dM6RHkUEZmaWrimtLsDMzEaXg97MLHEOejOzxDnozcwS56A3M0vcKa0uoJo5c+bEggULWl2GmdmE0dXV9W8R0VFt3rgM+gULFlAqlVpdhpnZhCHpx7Xm+dKNmVniHPRmZolz0JuZJc5Bb2aWOAe9mVni6j51I2kecD8wFwhga0TcNaCPgLuAtcBbwDUR8Xg272rg5qzr5oi4r3nlm6VpxxM9bNm1j5dePcqZM9u5fvX5rF/amXv5m3d0s33vQfoiKEhsXD6PF3rf5HvPHznZ56LzZgMMaus+9Bqvv913su20aQVuXb+oaj3V6iz9+MiIt5237S9/58NVv0YgV9vm9Yty7XMg93Fo5Jg1erzrUb3/XinpDOCMiHhc0gygC1gfEc9U9FkLfJ5y0C8H7oqI5ZJmAyWgSPlNogu4MCJ+NtQ2i8Vi+PFKm6x2PNHDjQ92c/TYe2Hb3lbg9ssW5frhv3lHN1/b85PRLJH2tgKfvLCTb3b19KtzCvDuqG65bO6MqbzyxjsjXv6qFfP7hX21fd42RSA41vdeRtY6Do0cs0aP9wmSuiKiWG1e3Us3EfHyibPziHgDeBYYuPV1wP1RtgeYmb1BrAZ2R8SRLNx3A2tyV242CW3Zta/fDz3A0WN9bNm1L9fy2/ceHI2y+jl6rI/tew8OqnMsQh5oKORh8D6qts+PvRv9Qh5qH4dGjlmjxzuPYV2jl7QAWArsHTCrE6jcc4eytlrt1da9SVJJUqm3t3c4ZZkl5aVXjw6rfaC+MRpjYqy2MxoG1p5339bq28gxa/R455E76CX9AvBN4A8j4vWmVZCJiK0RUYyIYkdH1b/iNZsUzpzZPqz2gQpSM8tp+XZGw8Da8+7bWn0bOWaNHu88cgW9pDbKIf+XEfFglS49wLyK6bOytlrtZlbD9avPp72t0K+tva1w8uZgPSduQI6m9rYCG5fPG1TnWD3GN3fG1IaWH7iPqu3ztimirdD/DaHWcWjkmDV6vPOoe1yyJ2r+DHg2Ir5Uo9tO4FMqWwG8FhEvA7uAVZJmSZoFrMrazKyG9Us7uf2yRXTObEdA58z2Yd2Y27x+EVetmH/yrLUgcdWK+SefYDnhovNmV207bVr/0DltWoE7r1wyqJ7N6xcNqvNLVy5paNt52/bedEnV7eRtG/jUTbV9vuWKxWy5fHGu49DIMWv0eOeR56mbi4F/BLp5717LHwHzASLinuzN4MuUb7S+Bfx2RJSy5T+d9Qe4LSL+vF5RfurGzGx4hnrqpu5z9BHxT8CQF+Oi/G7xuRrztgHbctRpZmajwH8Za2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJa7uwCOStgG/ARyOiF+pMv964Dcr1vcBoCMijkh6EXgD6AOO1xr9xMzMRk+eM/qvUh4isKqI2BIRSyJiCXAj8H8i4khFl49l8x3yZmYtUDfoI+JR4Ei9fpmNwPaGKjIzs6Zq2jV6Sf+O8pn/NyuaA/i2pC5Jm+osv0lSSVKpt7e3WWWZmU16zbwZ+5+A7w24bHNxRHwI+HXgc5I+UmvhiNgaEcWIKHZ0dDSxLDOzya2ZQb+BAZdtIqIn+3wYeAhY1sTtmZlZDk0Jekm/CHwU+JuKtlMlzTjxGlgF/KAZ2zMzs/zyPF65HVgJzJF0CLgFaAOIiHuybp8Avh0RP69YdC7wkKQT2/mriPi75pVuZmZ51A36iNiYo89XKT+GWdl2AFg80sLMzKw5/JexZmaJc9CbmSXOQW9mljgHvZlZ4hz0ZmaJc9CbmSXOQW9mljgHvZlZ4hz0ZmaJc9CbmSXOQW9mljgHvZlZ4hz0ZmaJc9CbmSXOQW9mljgHvZlZ4uoGvaRtkg5LqjoMoKSVkl6T9GT28cWKeWsk7ZO0X9INzSzczMzyyXNG/1VgTZ0+/xgRS7KPWwEkFYC7gV8HLgA2SrqgkWLNzGz46gZ9RDwKHBnBupcB+yPiQES8AzwArBvBeszMrAHNukb/YUlPSXpY0i9nbZ3AwYo+h7K2qiRtklSSVOrt7W1SWWZm1oygfxw4OyIWA/8T2DGSlUTE1ogoRkSxo6OjCWWZmRk0Iegj4vWIeDN7/S2gTdIcoAeYV9H1rKzNzMzGUMNBL+n9kpS9Xpat86fAY8BCSedImgpsAHY2uj0zMxueU+p1kLQdWAnMkXQIuAVoA4iIe4DLgd+TdBw4CmyIiACOS7oO2AUUgG0R8fSofBVmZlaTypk8vhSLxSiVSq0uw8xswpDUFRHFavP8l7FmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWuLpBL2mbpMOSflBj/m9K+r6kbkn/LGlxxbwXs/YnJfkfzJuZtUCeM/qvAmuGmP8C8NGIWAT8MbB1wPyPRcSSWv8Q38zMRlfdoQQj4lFJC4aY/88Vk3soDwJuZmbjRLOv0V8LPFwxHcC3JXVJ2jTUgpI2SSpJKvX29ja5LDOzyavuGX1ekj5GOegvrmi+OCJ6JJ0O7Jb0w4h4tNryEbGV7LJPsVgcfwPZmplNUE05o5f0QeBeYF1E/PREe0T0ZJ8PAw8By5qxPTMzy6/hoJc0H3gQ+K2I+FFF+6mSZpx4DawCqj65Y2Zmo6fupRtJ24GVwBxJh4BbgDaAiLgH+CLwPuBPJQEcz56wmQs8lLWdAvxVRPzdKHwNZmY2hDxP3WysM/8zwGeqtB8AFg9ewszMxpL/MtbMLHEOejOzxDnozcwS56A3M0ucg97MLHEOejOzxDnozcwS56A3M0ucg97MLHEOejOzxDnozcwS56A3M0ucg97MLHEOejOzxDnozcwS56A3M0tcrsHBJW0DfgM4HBG/UmW+gLuAtcBbwDUR8Xg272rg5qzr5oi4rxmF28Sx44ketuzax0uvHuXMme1cv/p81i/tzLXszTu62b73IH0RFCQ2Lp/H5vWLcm8HyLXtasuWfnyk6raX37abV9545+Syc2dMZe9Nlwxa5yVfeoTnDv/85PTC009l+bnvG7TOF3rf5HvPHznZ76LzZtP14s/4f31xsm16QfzwtrVV98fup/91UD03rr2gqfsi7/Gy8UkRUb+T9BHgTeD+GkG/Fvg85aBfDtwVEcslzQZKQBEIoAu4MCJ+NtT2isVilEql4X4tNg7teKKHGx/s5uixvpNt7W0Fbr9sUd3wuHlHN1/b85NB7VetmD8o7Kttp22KQHCsIjCrbbvasoUpou/dwT8b0wvqF8AnDAz7gSHfDKL8QzQSjeyLvMfLWktSVzaM6yC5Lt1ExKPAkSG6rKP8JhARsQeYKekMYDWwOyKOZOG+G1gzvPJtItuya1+/0AA4eqyPLbv21V12+96DudurbefYu9Ev2Gptu9qy1UIeqBryQL8zaqDpIQ8jD3lobF/kPV42fjXrGn0nUPnTdyhrq9U+iKRNkkqSSr29vU0qy1rtpVePDqu9Ul+N3zartedZX62+w1k2NXn3xWTeRykYNzdjI2JrRBQjotjR0dHqcqxJzpzZPqz2SgUpd3ue9dXqO5xlU5N3X0zmfZSCZgV9DzCvYvqsrK1Wu00S168+n/a2Qr+29rbCyZuDQ9m4fF7u9mrbaZsi2gr93xSqbbvasoUp1d9kpheqt8+dMbXf9MLTT63arxHVt5xPI/si7/Gy8atZQb8T+JTKVgCvRcTLwC5glaRZkmYBq7I2myTWL+3k9ssW0TmzHQGdM9tz39jbvH4RV62Yf/IMviBVvRFbaztbrljMlssX1912tWX/2xWLq277h7etHRTq1Z662f2FlYPCfuHpp1Zd50Xnze7X76LzZg96Q5leEC/c8fGqy1er584rlzRtX/hG7MSX96mb7cBKYA7wCnAL0AYQEfdkj1d+mfKN1reA346IUrbsp4E/ylZ1W0T8eb3t+akbM7PhGeqpm1zP0UfExjrzA/hcjXnbgG15tmNmZs03bm7GmpnZ6HDQm5klzkFvZpY4B72ZWeIc9GZmiXPQm5klzkFvZpY4B72ZWeIc9GZmiXPQm5klzkFvZpY4B72ZWeIc9GZmiXPQm5klzkFvZpY4B72ZWeJyBb2kNZL2Sdov6YYq8/+7pCezjx9JerViXl/FvJ3NLN7MzOqrO8KUpAJwN3AJcAh4TNLOiHjmRJ+I+C8V/T8PLK1YxdGIWNK8ks3MbDjynNEvA/ZHxIGIeAd4AFg3RP+NwPZmFGdmZo3LE/SdwMGK6UNZ2yCSzgbOAb5b0TxdUknSHknra21E0qasX6m3tzdHWWZmlkezb8ZuAL4REX0VbWdnI5P/Z+BOSedVWzAitkZEMSKKHR0dTS7LzGzyyhP0PcC8iumzsrZqNjDgsk1E9GSfDwCP0P/6vZmZjbI8Qf8YsFDSOZKmUg7zQU/PSPolYBbwfyvaZkmalr2eA1wEPDNwWTMzGz11n7qJiOOSrgN2AQVgW0Q8LelWoBQRJ0J/A/BARETF4h8AviLpXcpvKndUPq1jZmajT/1zeXwoFotRKpVaXYaZ2YQhqSu7HzqI/zLWzCxxDnozs8Q56M3MEuegNzNLnIPezCxxDnozs8Q56M3MEuegNzNLnIPezCxxDnozs8Q56M3MEuegNzNLnIPezCxxDnozs8Q56M3MEpcr6CWtkbRP0n5JN1SZf42kXklPZh+fqZh3taTnso+rm1m8mZnVV3eEKUkF4G7gEuAQ8JiknVVGivp6RFw3YNnZwC1AEQigK1v2Z02p3szM6spzRr8M2B8RByLiHeABYF3O9a8GdkfEkSzcdwNrRlaqmZmNRJ6g7wQOVkwfytoG+qSk70v6hqR5w1wWSZsklSSVent7c5RlZmZ5NOtm7N8CCyLig5TP2u8b7goiYmtEFCOi2NHR0aSyzMwsT9D3APMqps/K2k6KiJ9GxNvZ5L3AhXmXNTOz0ZUn6B8DFko6R9JUYAOws7KDpDMqJi8Fns1e7wJWSZolaRawKmszM7MxUvepm4g4Luk6ygFdALZFxNOSbgVKEbET+ANJlwLHgSPANdmyRyT9MeU3C4BbI+LIKHwdZmZWgyKi1TUMUiwWo1QqtboMM7MJQ1JXRBSrzfNfxpqZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZonLFfSS1kjaJ2m/pBuqzP+CpGckfV/S30s6u2Jen6Qns4+dA5c1M7PRVXcoQUkF4G7gEuAQ8JiknRHxTEW3J4BiRLwl6feAPwGuzOYdjYglTa7bzMxyynNGvwzYHxEHIuId4AFgXWWHiPiHiHgrm9wDnNXcMs3MbKTyBH0ncLBi+lDWVsu1wMMV09MllSTtkbS+1kKSNmX9Sr29vTnKMjOzPOpeuhkOSVcBReCjFc1nR0SPpHOB70rqjojnBy4bEVuBrVAeHLyZdZmZTWZ5zuh7gHkV02dlbf1I+jXgJuDSiHj7RHtE9GSfDwCPAEsbqNfMzIYpT9A/BiyUdI6kqcAGoN/TM5KWAl+hHPKHK9pnSZqWvZ4DXARU3sQ1M7NRVvfSTUQcl3QdsAsoANsi4mlJtwKliNgJbAF+AfhrSQA/iYhLgQ8AX5H0LuU3lTsGPK1jZmajTBHj73J4sViMUqnU6jLMzCYMSV0RUaw2z38Za2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJc5Bb2aWOAe9mVniHPRmZolz0JuZJS7X4OCS1gB3UR5h6t6IuGPA/GnA/cCFwE+BKyPixWzejcC1QB/wBxGxq2nVV9jxRA9bdu3jpVePcubMdq5ffT5ArrYv7ujm9bf7Tq7rtGmFftMnvHjHxznnhv9N5VAtAk6fMZVX3njnZNvcGVM5rb2N5w7//GTbwtNPZfcXVg5a5807utm+9yB9ERQkNi6fx+b1ixrbGWZmFeqOMCWpAPwIuAQ4RHkM2Y2VQwJK+n3ggxHxWUkbgE9ExJWSLgC2A8uAM4HvAP8+IganaIXhjjC144kebnywm6PH3ltt2xSB4FhfDNk2lgaG/c07uvnanp8M6nfVivkOezMblkZHmFoG7I+IAxHxDvAAsG5An3XAfdnrbwD/UeXBY9cBD0TE2xHxArA/W19Tbdm1r1/IAxx7NwYFerW2sVR5hg+wfe/Bqv1qtZuZjUSeoO8EKpPnUNZWtU9EHAdeA96Xc1kAJG2SVJJU6u3tzVd95qVXjw6r/3jRV+O3qVrtZmYjMW5uxkbE1ogoRkSxo6NjWMueObN9lKoaXQVpWO1mZiORJ+h7gHkV02dlbVX7SDoF+EXKN2XzLNuw61efT3tboV9b2xTRVlDdtrG08PRT+01vXD6var9a7WZmI5En6B8DFko6R9JUYAOwc0CfncDV2evLge9G+S7vTmCDpGmSzgEWAv/SnNLfs35pJ7dftojOme0I6JzZzpYrFrPl8sV12+68cgmnTev/JjFw+oQX7/g4A98mRPkpm0pzZ0wdFOrVnrrZvH4RV62Yf/IMviD5RqyZNV3dp24AJK0F7qT8eOW2iLhN0q1AKSJ2SpoO/AWwFDgCbIiIA9myNwGfBo4DfxgRD9fb3nCfujEzm+yGeuomV9CPNQe9mdnwNPp4pZmZTWAOejOzxDnozcwS56A3M0vcuLwZK6kX+HGTVzsH+Lcmr3O0ueaxMxHrds1jY6LUfHZEVP1r03EZ9KNBUqnWHenxyjWPnYlYt2seGxOx5oF86cbMLHEOejOzxE2moN/a6gJGwDWPnYlYt2seGxOx5n4mzTV6M7PJajKd0ZuZTUoOejOzxCUb9JKukPS0pHcl1Xw0StKLkrolPSmppf9JbRg1r5G0T9J+STeMZY1Vapktabek57LPs2r068v28ZOSBv6b6zFRb79l/07769n8vZIWjH2Vg+Wo+xpJvRX79zOtqLOinm2SDkv6QY35kvQ/sq/n+5I+NNY1VpOj7pWSXqvYz18c6xpHLCKS/AA+AJwPPAIUh+j3IjCn1fXmrZnyv4p+HjgXmAo8BVzQwpr/BLghe30D8F9r9Huzxfu27n4Dfh+4J3u9Afj6OPieyFP3NcCXW11rRT0fAT4E/KDG/LXAw5SHc1gB7G11zTnrXgn8r1bXOZKPZM/oI+LZiNjX6jqGI2fNeQZrH0uVA8PfB6xvYS1DaWSQ+1Yab8e7roh4lPK4FLWsA+6Psj3ATElnjE11teWoe8JKNuiHIYBvS+qStKnVxeSQe8D1MTI3Il7OXv8rMLdGv+nZ4O97JLXizaCRQe5bKe/x/mR2GeQbksb7WJTj7Xt4OD4s6SlJD0v65VYXk9cprS6gEZK+A7y/yqybIuJvcq7m4ojokXQ6sFvSD7N39lHRpJrH1FA1V05EREiq9bzu2dl+Phf4rqTuiHi+2bVOUn8LbI+ItyX9LuXfSv5Di2tK0eOUv4/fzEbd20F5eNRxb0IHfUT8WhPW0ZN9PizpIcq/Ko9a0Deh5jEZcL3SUDVLekXSGRHxcvbr9+Ea6zixnw9IeoTysJNjGfTDGeT+0IBB7lupbt0RUVnjvZTvm4xnY/493AwR8XrF629J+lNJcyJi3P/Ds0l96UbSqZJmnHgNrAKq3nEfR/IM1j6WKgeGvxoY9FuJpFmSpmWv5wAXAc+MWYVljQxy30p16x5wfftS4NkxrG8kdgKfyp6+WQG8VnH5b9yS9P4T92wkLaOcn60+Ecin1XeDR+sD+ATla39vA68Au7L2M4FvZa/PpfwUw1PA05Qvn4zrmrPptcCPKJ8Rt7rm9wF/DzwHfAeYnbUXgXuz178KdGf7uRu4tkW1DtpvwK3Apdnr6cBfA/uBfwHObeW+HUbdt2ffv08B/wD8Uovr3Q68DBzLvp+vBT4LfDabL+Du7OvpZoin4sZZ3ddV7Oc9wK+2uua8H/4XCGZmiZvUl27MzCYDB72ZWeIc9GZmiXPQm5klzkFvZpY4B72ZWeIc9GZmifv/MjJ3XlmtpN0AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_classes = 3\n",
        "num_features = 1\n",
        "\n",
        "# Training parameters.\n",
        "learning_rate = 0.01\n",
        "training_steps = 100\n",
        "batch_size = 40\n",
        "display_step = 20\n",
        "\n",
        "train_data=tf.data.Dataset.from_tensor_slices((train_x1,train_y))\n",
        "train_data=train_data.repeat().shuffle(50).batch(batch_size).prefetch(1)"
      ],
      "metadata": {
        "id": "yi1JFjDdFfal"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "W = tf.Variable(tf.ones([num_features, num_classes]), name=\"weight\")\n",
        "b = tf.Variable(tf.zeros([num_classes]), name=\"bias\")"
      ],
      "metadata": {
        "id": "yLbchBzcFhyl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Logistic regression (Wx + b).\n",
        "def logistic_regression(x):\n",
        "    return tf.nn.softmax(tf.matmul(x, W) + b)\n",
        "\n",
        "# Cross-Entropy loss function.\n",
        "def cross_entropy(y_pred, y_true):\n",
        "    y_true = tf.one_hot(y_true, depth=num_classes)\n",
        "    y_pred = tf.clip_by_value(y_pred, 1e-9, 1.)\n",
        "\n",
        "    return tf.reduce_mean(-tf.reduce_sum(y_true * tf.math.log(y_pred)))\n",
        "\n",
        "# Accuracy metric.\n",
        "def accuracy(y_pred, y_true):\n",
        "  correct_prediction = tf.equal(tf.argmax(y_pred, 1), tf.cast(y_true, tf.int64))\n",
        "\n",
        "  return tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
        "\n",
        "# Stochastic gradient descent optimizer.\n",
        "optimizer = tf.optimizers.SGD(learning_rate)"
      ],
      "metadata": {
        "id": "sEa7avLKFk7C"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def run_optimization(x, y):\n",
        "    with tf.GradientTape() as g:\n",
        "        pred = logistic_regression(x)\n",
        "        loss = cross_entropy(pred, y)\n",
        "\n",
        "    gradients = g.gradient(loss, [W, b])\n",
        "  \n",
        "    optimizer.apply_gradients(zip(gradients, [W, b]))"
      ],
      "metadata": {
        "id": "9moXtFrQFoI1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for (step, (batch_x, batch_y)) in enumerate(train_data.take(training_steps), 1):\n",
        "  run_optimization(tf.cast(batch_x, tf.float32), batch_y)\n",
        "\n",
        "\n",
        "  if step % display_step == 0:\n",
        "    pred = logistic_regression(tf.cast(batch_x, tf.float32))\n",
        "    loss = cross_entropy(pred, batch_y)\n",
        "    acc = accuracy(pred, batch_y)\n",
        "    print(\"step: %i, loss: %f, accuracy: %f\" % (step, loss, acc))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PZHvYhrOGTjz",
        "outputId": "d96be0bf-ab28-42cc-9fb6-57f83e16ca61"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "step: 20, loss: 17.932350, accuracy: 0.850000\n",
            "step: 40, loss: 18.893982, accuracy: 0.925000\n",
            "step: 60, loss: 14.597206, accuracy: 0.950000\n",
            "step: 80, loss: 11.995378, accuracy: 0.950000\n",
            "step: 100, loss: 13.934515, accuracy: 0.975000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred = logistic_regression(tf.cast(test_x1, tf.float32))\n",
        "\n",
        "print(\"Test Accuracy: %f\" % accuracy(pred, test_y))\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TOv9Yl_yFtwO",
        "outputId": "9ea96ba5-b8dd-4bae-d32f-b65dd7b7cd95"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test Accuracy: 0.933333\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(pred, test_y, 'o')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "HSvp47vJFzBq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "5de8a966-c8e9-407f-ea40-a7e891bdae1d"
      },
      "execution_count": null,
      "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"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Our model was able to produce approximately 93% accuracy in correctly classifying the three different flowers based on the chosen criteria."
      ],
      "metadata": {
        "id": "Ob-w9T5eaZUV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Implementing Support Vector Machine**"
      ],
      "metadata": {
        "id": "hMOxvlDTXYav"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "We will use the iris data set, classifying flowers by the feature petal width and sepal length"
      ],
      "metadata": {
        "id": "o8aaA-QacNIJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()\n",
        "from tensorflow._api.v2 import data\n",
        "import numpy as np\n",
        "\n",
        "\n",
        "sess = tf.Session()\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "svm_x = np.array([[x[0], x[3]] for x in iris.data])\n",
        "svm_y = np.array([1 if y==0 else -1 for y in iris.target])"
      ],
      "metadata": {
        "id": "f0G-zcDOXBYB"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train = np.random.choice(len(svm_x), round(len(svm_x)*0.8), replace=False)\n",
        "test = np.array(list(set(range(len(svm_x)))-set(train)))\n",
        "svm_trainx = svm_x[train]\n",
        "svm_testx = svm_x[test]\n",
        "svm_trainy = svm_y[train]\n",
        "svm_testy = svm_y[test]"
      ],
      "metadata": {
        "id": "OgNHsNxUgskb"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size=100\n",
        "x_data = tf.placeholder(shape=[None,2], dtype=tf.float32)\n",
        "y_data = tf.placeholder(shape=[None,1], dtype=tf.float32)\n",
        "W = tf.Variable(tf.random_normal(shape=[2,1]))\n",
        "b = tf.Variable(tf.random_normal(shape=[1,1]))\n",
        "\n",
        "\n",
        "model_output = tf.subtract(tf.matmul(x_data,W), b)\n"
      ],
      "metadata": {
        "id": "gjAIXJvBmRkb"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "l2_norm = tf.reduce_sum(tf.square(W))\n",
        "alpha = 0.1\n",
        "classification_term = tf.reduce_mean(tf.maximum(0., tf.subtract(1., tf.multiply(model_output,y_data))))\n",
        "loss = tf.add(tf.multiply(alpha,l2_norm), classification_term)\n",
        "\n"
      ],
      "metadata": {
        "id": "yA80ltKHrdIy"
      },
      "execution_count": 81,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "prediction = tf.sign(model_output)\n",
        "accuracy = tf.reduce_mean(tf.cast(tf.equal(prediction, y_data), tf.float32))\n",
        "residuals = prediction - y_data"
      ],
      "metadata": {
        "id": "Vdr8qV-z9pnI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "svm_opt = tf.train.GradientDescentOptimizer(0.01)\n",
        "train_step = svm_opt.minimize(loss)\n",
        "init = tf.global_variables_initializer()\n",
        "sess.run(init)"
      ],
      "metadata": {
        "id": "0jLIUNr0_ZSg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with tf.name_scope('Loss'):\n",
        "  tf.summary.histogram('Histogram_Errors', accuracy)\n",
        "  tf.summary.histogram('Histogram_Residuals', residuals)\n",
        "  Loss_summary_OP = tf.summary.merge_all()"
      ],
      "metadata": {
        "id": "jMCcLts-CEq2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "loss_vec = []\n",
        "train_accuracy = []\n",
        "test_accuracy = []"
      ],
      "metadata": {
        "id": "XTY5UD-G_Zij"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorboard.summary._tf import summary\n",
        "for i in range(5000):\n",
        "  rand_index = np.random.choice(len(svm_trainx), size=batch_size)\n",
        "  rand_x = svm_trainx[rand_index]\n",
        "  rand_y = np.transpose([svm_trainy[rand_index]])\n"
      ],
      "metadata": {
        "id": "GqOoXhR6_Zus"
      },
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Support Vector Regression**"
      ],
      "metadata": {
        "id": "EjHmZ1PlXB-6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()\n",
        "from sklearn import datasets\n",
        "from tensorflow.python.framework import ops\n",
        "ops.reset_default_graph()"
      ],
      "metadata": {
        "id": "6Que00728A-U"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Load the data and creaye datasets"
      ],
      "metadata": {
        "id": "9XvPC_OB8xzB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create graph\n",
        "sess = tf.Session()\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "x_vals = np.array([x[3] for x in iris.data])\n",
        "y_vals = np.array([y[0] for y in iris.data])\n",
        "\n",
        "train_indices = np.random.choice(len(x_vals), round(len(x_vals)*0.8), replace=False)\n",
        "test_indices = np.array(list(set(range(len(x_vals))) - set(train_indices)))\n",
        "x_vals_train = x_vals[train_indices]\n",
        "x_vals_test = x_vals[test_indices]\n",
        "y_vals_train = y_vals[train_indices]\n",
        "y_vals_test = y_vals[test_indices]"
      ],
      "metadata": {
        "id": "L_z35GoN8BSJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Declare batch size\n",
        "batch_size = 50\n",
        "\n",
        "# Initialize placeholders\n",
        "x_data = tf.placeholder(shape=[None, 1], dtype=tf.float32)\n",
        "y_target = tf.placeholder(shape=[None, 1], dtype=tf.float32)\n",
        "\n",
        "# Create variables for linear regression\n",
        "W = tf.Variable(tf.random_normal(shape=[1,1]))\n",
        "b = tf.Variable(tf.random_normal(shape=[1,1]))\n",
        "\n",
        "model_output = tf.add(tf.matmul(x_data, W), b)\n",
        "epsilon = tf.constant([0.5])\n",
        "loss = tf.reduce_mean(tf.maximum(0., tf.subtract(tf.abs(tf.subtract(model_output, y_target)), epsilon)))\n",
        "my_opt = tf.train.GradientDescentOptimizer(0.075)\n",
        "train_step = my_opt.minimize(loss)\n",
        "\n",
        "init = tf.global_variables_initializer()\n",
        "sess.run(init)"
      ],
      "metadata": {
        "id": "f_kV-AEb8BgH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Training loop\n",
        "train_loss = []\n",
        "test_loss = []\n",
        "for i in range(200):\n",
        "    rand_index = np.random.choice(len(x_vals_train), size=batch_size)\n",
        "    rand_x = np.transpose([x_vals_train[rand_index]])\n",
        "    rand_y = np.transpose([y_vals_train[rand_index]])\n",
        "    sess.run(train_step, feed_dict={x_data: rand_x, y_target: rand_y})\n",
        "    \n",
        "    temp_train_loss = sess.run(loss, feed_dict={x_data: np.transpose([x_vals_train]), y_target: np.transpose([y_vals_train])})\n",
        "    train_loss.append(temp_train_loss)\n",
        "    \n",
        "    temp_test_loss = sess.run(loss, feed_dict={x_data: np.transpose([x_vals_test]), y_target: np.transpose([y_vals_test])})\n",
        "    test_loss.append(temp_test_loss)\n",
        "    if (i+1)%50==0:\n",
        "        print( str(i+1))\n",
        "        print('W = ' + str(sess.run(W)) + ' b = ' + str(sess.run(b)))\n",
        "        "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3Etx5eGK8Btk",
        "outputId": "a7382783-89b3-417c-e894-5970e1499e85"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "50\n",
            "W = [[2.67424]] b = [[2.0404081]]\n",
            "100\n",
            "W = [[1.9824401]] b = [[3.1189072]]\n",
            "150\n",
            "W = [[1.43434]] b = [[3.9589076]]\n",
            "200\n",
            "W = [[1.14769]] b = [[4.4224095]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "[[slope]] = sess.run(W)\n",
        "[[y_intercept]] = sess.run(b)\n",
        "width = sess.run(epsilon)\n",
        "\n",
        "# Get best fit line\n",
        "best_fit = []\n",
        "best_fit_upper = []\n",
        "best_fit_lower = []\n",
        "for i in x_vals:\n",
        "    best_fit.append(slope*i+y_intercept)\n",
        "    best_fit_upper.append(slope*i+y_intercept+width)\n",
        "    best_fit_lower.append(slope*i+y_intercept-width)"
      ],
      "metadata": {
        "id": "8N-lyl-Y-wj5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(x_vals, y_vals, 'o', label='Data Points')\n",
        "plt.plot(x_vals, best_fit, 'r-', label='SVM Regression Line', linewidth=3)\n",
        "plt.plot(x_vals, best_fit_upper, 'r--', linewidth=2)\n",
        "plt.plot(x_vals, best_fit_lower, 'r--', linewidth=2)\n",
        "plt.ylim([0, 10])\n",
        "plt.legend(loc='lower right')\n",
        "plt.title('Sepal Length vs Petal Width')\n",
        "plt.xlabel('Petal Width')\n",
        "plt.ylabel('Sepal Length')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "3WqTdk4c-ww6",
        "outputId": "bef3391c-3d7e-4bda-c3a9-f11708ff8d26"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Implementing XOR Neural Network**"
      ],
      "metadata": {
        "id": "KN_hywZlRR9U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "in_nodes=2\n",
        "hd_nodes = 2\n",
        "out_node = 1\n",
        "W_hd = np.random.uniform(size=(in_nodes,hd_nodes))\n",
        "b_hd =np.random.uniform(size=(1,hd_nodes))\n",
        "W_out = np.random.uniform(size=(hd_nodes,out_node))\n",
        "b_out = np.random.uniform(size=(1,out_node))\n",
        "\n",
        "learning_rate=0.5\n",
        "epochs = [200, 500, 1000, 5000, 10000]\n"
      ],
      "metadata": {
        "id": "37l3RfeERVWu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "inputs = np.array([[0,0],[0,1],[1,0],[1,1]])\n",
        "expected_output = np.array([[0],[1],[1],[0]])"
      ],
      "metadata": {
        "id": "C1oeEk3yRd6r"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Create sigmoid function that will be used for the hidden and output layers, as well as the derivative for the back propogation."
      ],
      "metadata": {
        "id": "RunricVjat_h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def sigmoid (x):\n",
        "    return 1/(1 + np.exp(-x))\n",
        "\n",
        "\n",
        "def sigmoid_der(x):\n",
        "    return x * (1 - x)\n"
      ],
      "metadata": {
        "id": "xB6KBDvYReBs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Model implementation for a single node"
      ],
      "metadata": {
        "id": "yDklWhBva-BA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for i in epochs:\n",
        "  for x in range(epochs[epochs.index(i)]):\n",
        "    hd_activation = np.dot(inputs,W_hd) + b_hd\n",
        "    out_hd = sigmoid(hd_activation)\n",
        "    out_activation = np.dot(out_hd,W_out) + b_out\n",
        "    out_pred = sigmoid(out_activation)\n",
        "\n",
        "    #Backpropagation\n",
        "    error = expected_output - out_pred\n",
        "    d_out_pred = error * sigmoid_der(out_pred)\n",
        "    error_hd = d_out_pred.dot(W_out.T)\n",
        "    d_hd = error_hd * sigmoid_der(out_hd)\n",
        "\n",
        "    #Updating Weights and Biases\n",
        "    W_out += out_hd.T.dot(d_out_pred) * learning_rate\n",
        "    b_out += np.sum(d_out_pred,axis=0,keepdims=True) * learning_rate\n",
        "    W_hd += inputs.T.dot(d_hd) * learning_rate\n",
        "    b_hd += np.sum(d_hd,axis=0,keepdims=True) * learning_rate\n",
        "\n",
        "  print(\"Epochs:\" + str(epochs[epochs.index(i)]))\n",
        "  print(\"Expected: \\n\" + str(expected_output) + \"\\n Predicted: \\n\" + str(out_pred))\n",
        "  print(\"\\n\")\n"
      ],
      "metadata": {
        "id": "EDeb6eoDReN_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "664b95ad-9efb-4575-d31b-5dfdf7893b4c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epochs:200\n",
            "Expected: \n",
            "[[0]\n",
            " [1]\n",
            " [1]\n",
            " [0]]\n",
            " Predicted: \n",
            "[[0.03000045]\n",
            " [0.97361219]\n",
            " [0.97363211]\n",
            " [0.0276692 ]]\n",
            "\n",
            "\n",
            "Epochs:500\n",
            "Expected: \n",
            "[[0]\n",
            " [1]\n",
            " [1]\n",
            " [0]]\n",
            " Predicted: \n",
            "[[0.02813081]\n",
            " [0.9753343 ]\n",
            " [0.97535224]\n",
            " [0.02581013]]\n",
            "\n",
            "\n",
            "Epochs:1000\n",
            "Expected: \n",
            "[[0]\n",
            " [1]\n",
            " [1]\n",
            " [0]]\n",
            " Predicted: \n",
            "[[0.02522989]\n",
            " [0.97798495]\n",
            " [0.97799997]\n",
            " [0.02296271]]\n",
            "\n",
            "\n",
            "Epochs:5000\n",
            "Expected: \n",
            "[[0]\n",
            " [1]\n",
            " [1]\n",
            " [0]]\n",
            " Predicted: \n",
            "[[0.01786668]\n",
            " [0.98460034]\n",
            " [0.98460892]\n",
            " [0.01593102]]\n",
            "\n",
            "\n",
            "Epochs:10000\n",
            "Expected: \n",
            "[[0]\n",
            " [1]\n",
            " [1]\n",
            " [0]]\n",
            " Predicted: \n",
            "[[0.01258574]\n",
            " [0.98924936]\n",
            " [0.98925424]\n",
            " [0.01105428]]\n",
            "\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Our model is accurately able to implement the XOR algorithm. We also observe that increasing the value used for epochs increases the model's accuracy"
      ],
      "metadata": {
        "id": "VsUEcgl8JMpG"
      }
    }
  ]
}