{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5d978533-e16a-47ca-8a6b-068096157b51",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Weights at the beginning: \n",
      "[0.05570723 0.91606935 0.03272122 0.4935642 ]\n",
      "Training finished. Final weights:\n",
      "[0.06978601 0.83112226 0.03200317 0.56776683]\n"
     ]
    }
   ],
   "source": [
    "#Imports\n",
    "import numpy as np\n",
    "import random\n",
    "\n",
    "#Variables\n",
    "eta = 0.01 # Learning Rate\n",
    "n = 3 # Dimension, \n",
    "training_set_size = 12 #Size of the training set\n",
    "\n",
    "#Initialise arrays\n",
    "x = np.mgrid[0.0:training_set_size, 0:n+1] # Training set, only use x[0] subarray\n",
    "w = np.arange(0, n + 1 , 1.0) # Weight vector\n",
    "t = np.arange(0, training_set_size, 1)\n",
    "\n",
    "# To ensure that random numbers stay the same for multiple executions\n",
    "random.seed(5)\n",
    "\n",
    "#Fill arrays with values\n",
    "for i in range(training_set_size):\n",
    "    #Set value of t either to +1 or -1\n",
    "    random_number = random.random() #Call random number\n",
    "    if(random_number < 0.5):\n",
    "        t[i] = -1\n",
    "    else:\n",
    "        t[i] = +1\n",
    "        \n",
    "    #Set values of x, take random values between -5 and 5\n",
    "    for j in range(n + 1):\n",
    "        random_number_2 = random.random()\n",
    "        x[0][i][j] = (random.random())*5.0\n",
    "        if(random_number_2 < 0.5):\n",
    "            x[0][i][j] *= -1\n",
    "            \n",
    "\n",
    "# Initialise weight vector, weights between 0 and +1\n",
    "for i in range(n + 1):\n",
    "    w[i] = float(random.random())\n",
    "\n",
    "print(\"Weights at the beginning: \")\n",
    "print(w)\n",
    "\n",
    "# Training loop\n",
    "termination = False # Python has no do-while loop\n",
    "\n",
    "while(termination == False):\n",
    "    # Select random training example\n",
    "    random_number = int(random.random()*training_set_size)\n",
    "    \n",
    "    #Calculate Dot product\n",
    "    result = 0.0\n",
    "    for i in range(n+1):\n",
    "        result += w[i]*x[0][random_number][i]\n",
    "        \n",
    "    # Check termination condition\n",
    "    if(((np.sign(result) == -1) and (t[random_number] == -1)) or ((np.sign(result) == 1) and (t[random_number] == 1))):\n",
    "        termination = True\n",
    "        break\n",
    "    \n",
    "    # If no termination, update weights\n",
    "    else:\n",
    "        for i in range(n+1):\n",
    "            w[i] += eta*(t[random_number] - np.sign(result)) * x[0][random_number][i]\n",
    "\n",
    "print(\"Training finished. Final weights:\")\n",
    "print(w)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c125185c-4cf3-4e43-83c7-9e1c3936e268",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}