The Recommending Agricultural Product Sales Promotion Mode in E-Commerce Using Reinforcement Learning with Contextual Multiarmed Bandit Algorithms

Figure 6

Hybrid-LinUCB, UCB, and A/B testing have 80% selection probability, number of attempts, and recommendation probability.