thompson_sampling.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "362ae951",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "718724d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv(r\"C:\\Users\\Lenovo\\Desktop\\450 in 62\\Ads_CTR_Optimisation.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "fce5aadc",
   "metadata": {},
   "outputs": [],
   "source": [
    "N =10000\n",
    "d = 10\n",
    "ads_selected = []\n",
    "no_of_reward_1 = [0]*d\n",
    "no_of_reward_0 = [0]*d\n",
    "total_reward = 0\n",
    "for n in range(0,N):\n",
    "    ad = 0\n",
    "    Max_random = 0\n",
    "    for i in range(0,d):\n",
    "        random_beta = random.betavariate(no_of_reward_1[i] + 1 ,no_of_reward_0[i] + 1)\n",
    "        if(random_beta > Max_random):\n",
    "            Max_random = random_beta\n",
    "            ad = i\n",
    "    ads_selected.append(ad)\n",
    "    reward = dataset.values[n,ad]\n",
    "    if reward == 1:\n",
    "        no_of_reward_1[ad] += 1\n",
    "    else:\n",
    "        no_of_reward_0[ad] += 1\n",
    "    total_reward += reward\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "46a7cc8a",
   "metadata": {},
   "outputs": [
    {
     "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.hist(ads_selected)\n",
    "plt.title(\"Thompson sampling\")\n",
    "plt.xlabel(\"No of Ads\")\n",
    "plt.ylabel(\"CTR\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2b00407d",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}