Recommendation System Design for Social Media using Reinforcement Learning
Xiangyu Gao, Meikang Qiu · 2022
Recommendation System plays an important role in capturing consumers’ preference. In order to better utilize the effectiveness of the recommendation system, many social media platforms have developed algorithms to improve their performance. However, there are still a lot of platforms that keep recommending the same but useless contents to the customers, which even reduce the customers’ interest in using these social media platforms. In response to this problem, we propose to transplant reinforcement learning to build efficient and effective recommendation system. Specifically, we decide to update our recommendation system frequently by taking all customers’ behaviors into consideration to improve accuracy of the system’s recommendation output. The experiment results show that our system can gain more than 100% profits and can converge to the optimal result quickly.