task2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Optimization algorithms with default recommended parameters, additionaly each algorithm has an option to include interactions between variables\n",
    "1. IWLS\n",
    "2. SGD\n",
    "3. ADAM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\prusak.patryk\\AppData\\Local\\Temp\\ipykernel_28260\\1760283076.py:3: DeprecationWarning: \n",
      "Pyarrow will become a required dependency of pandas in the next major release of pandas (pandas 3.0),\n",
      "(to allow more performant data types, such as the Arrow string type, and better interoperability with other libraries)\n",
      "but was not found to be installed on your system.\n",
      "If this would cause problems for you,\n",
      "please provide us feedback at https://github.com/pandas-dev/pandas/issues/54466\n",
      "        \n",
      "  import pandas as pd\n"
     ]
    }
   ],
   "source": [
    "from logisticRegression import LogisticRegression, Optimizer\n",
    "from scipy.io.arff import loadarff\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def addInteractions(X):\n",
    "    newX=X\n",
    "    for i in range(X.shape[1]):\n",
    "        for j in range(i+1,X.shape[1]):\n",
    "            newX=np.c_[newX,np.multiply(X[:,i],X[:,j])]\n",
    "    return newX"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IWLS Converged after 6 iterations\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.728744939271255"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "logisticRegressor=LogisticRegression(noOfIterations=500, learningRate=0.001, optimizer=Optimizer.IWLS)\n",
    "raw_data = loadarff(\"./Datasets/blood-transfusion-service-center.arff\")\n",
    "df_data = pd.DataFrame(raw_data[0])\n",
    "df_no_class=df_data.drop('Class',axis=1)\n",
    "X = df_no_class.to_numpy()\n",
    "\n",
    "\n",
    "y = df_data['Class'].to_numpy()\n",
    "y[y==b'2']=1\n",
    "y[y==b'1']=0\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)\n",
    "costs = logisticRegressor.fit(X_train.astype(float),y_train.astype(float))\n",
    "y_pred=logisticRegressor.predict(X_test)\n",
    "\n",
    "sns.lineplot(costs)\n",
    "1-np.sum(np.abs(y_pred-y_test))/len(y_test)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adam Converged after 362 iterations\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.736842105263158"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "logisticRegressor=LogisticRegression(noOfIterations=500, learningRate=0.001, optimizer=Optimizer.Adam)\n",
    "raw_data = loadarff(\"./Datasets/blood-transfusion-service-center.arff\")\n",
    "df_data = pd.DataFrame(raw_data[0])\n",
    "df_no_class=df_data.drop('Class',axis=1)\n",
    "X = df_no_class.to_numpy()\n",
    "\n",
    "\n",
    "y = df_data['Class'].to_numpy()\n",
    "y[y==b'2']=1\n",
    "y[y==b'1']=0\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)\n",
    "costs = logisticRegressor.fit(X_train.astype(float),y_train.astype(float))\n",
    "y_pred=logisticRegressor.predict(X_test)\n",
    "\n",
    "sns.lineplot(costs)\n",
    "1-np.sum(np.abs(y_pred-y_test))/len(y_test)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SGD Converged after 54 iterations\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.7327935222672064"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "logisticRegressor=LogisticRegression(noOfIterations=500, learningRate=0.00000001, optimizer=Optimizer.SGD)\n",
    "raw_data = loadarff(\"./Datasets/blood-transfusion-service-center.arff\")\n",
    "df_data = pd.DataFrame(raw_data[0])\n",
    "df_no_class=df_data.drop('Class',axis=1)\n",
    "X = df_no_class.to_numpy()\n",
    "\n",
    "\n",
    "y = df_data['Class'].to_numpy()\n",
    "y[y==b'2']=1\n",
    "y[y==b'1']=0\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)\n",
    "costs = logisticRegressor.fit(X_train.astype(float),y_train.astype(float))\n",
    "y_pred=logisticRegressor.predict(X_test)\n",
    "\n",
    "sns.lineplot(costs)\n",
    "1-np.sum(np.abs(y_pred-y_test))/len(y_test)\n",
    "\n"
   ]
  }
 ],
 "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": "3.12.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}