MLR section 28 Thompson Sampling

Thompson Sampling - Reinforced learning

Importing the dataset

dataset <-  read.csv('G:\\RStudio\\udemy\\ml\\Machine Learning AZ\\Part 6 - Reinforcement Learning\\Section 33 - Thompson Sampling\\Thompson_Sampling\\Ads_CTR_Optimisation.csv')
head(dataset)

Implementing Thompson Sampling

There is package to use UCB. So we will create the algorithm from scatch

# initialize variables we will be using
N <-  10000
d <-  10
ads_selected <-  integer(0)
total_reward <-  0
# new parameters
numbers_of_rewards_1 <- integer(d)
numbers_of_rewards_0 <- integer(d)
for (n in 1:N) {
  
  ad <-  0
  max_random <-  0
    for ( i in 1 : d) { 
    random_beta <-  rbeta(n=1,shape1 = numbers_of_rewards_1[i] +1, 
                          shape2 = numbers_of_rewards_0[i]+1) 
    if (random_beta > max_random) {
      max_random <-  random_beta
      ad <-  i
    }
  }
  ads_selected <-  append(ads_selected, ad)
  reward <-  dataset[n,ad]
  if (reward ==1) {
      numbers_of_rewards_1[ad] <-  numbers_of_rewards_1[ad] +1
  } else {
    numbers_of_rewards_0[ad] <-  numbers_of_rewards_0[ad] +1
  }
  total_reward <-  total_reward + reward
  
}

Visualising the resuls by using a histogram

hist(ads_selected, 
     col = "Blue",
     main = "Histogram of Ad Selections",
     xlab = "Ads",
     ylab = "Number of times each ad was selected")