Thompson抽样算法-R

楔子：

Thompson抽样算法.PNG 贝叶斯推理.PNG

1、数据预处理

代码：

# Thompson Sampling

# Importing the dataset
dataset = read.csv('Ads_CTR_Optimisation.csv')

2、数据初始化

代码：

# Implementing Thompson Sampling
N = 10000
d = 10
ads_selected = integer(0)
numbers_of_rewards_1 = integer(d)
numbers_of_rewards_0 = integer(d)
total_reward = 0

3、ThompsonSampling

代码：

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
}

4、数据可视化

代码：

# Visualising the results
hist(ads_selected,
     col = 'blue',
     main = 'Histogram of ads selections',
     xlab = 'Ads',
     ylab = 'Number of times each ad was selected')

5、结果

最佳广告.PNG 点击次数.PNG

我们可以看出：无论是点击次数还是寻找最佳广告的速度，Thompson抽样算法都更胜一筹。