Optimizing Click-Through Rates in Online Advertising Using Thompson Sampling

School of Information Technology, Shanghai Jian Qiao University, Shanghai, China

^* Corresponding author: 2120350@stu.gench.edu.cn

Abstract

Advertisers must optimize ad selection to increase click-through rates (CTR) in an unpredictable environment as a result of the quick expansion of online advertising. Despite its effectiveness, traditional A/B testing is inefficient when it comes to dynamic user behavior and an ever- changing ad pool. This study integrates user behavior and ad context data using the TS algorithm to optimize ad selection through dynamic prediction of ad click-through rates. This allows for well-informed ad choices even when data is sparse and ad performance fluctuates. Although there were some missing pieces in the dataset utilized for the experiment, TS selected the Ad5 ad about 200 times in 200 trials. This suggests that the TS method can continue to achieve high accuracy and robustness in selection while optimizing the click-through rate in the presence of missing data. As a result, TS can adjust to complex data settings and performs better in advertising optimization. The study's findings indicate that the TS algorithm can give advertisers a practical tool for ad selection, allowing them to optimize their marketing efforts in a dynamic context.