Banner Ranking Based On Click Prediction In The E-Commerce

This paper presents an alternative banner component ranking model to improve personalized page design in e-commerce. It utilizes a machine learning model to predict the click probability of different banner types. This is achieved by incorporating factors such as user history, category and brand tendencies, and meta-features related to the banner components. The study reveals a 4.51% lift in conversion rates compared to the previous algorithm, showcasing the ML models’ effectiveness in improving home page banner click-through rates (CTR). The results underscore the value of personalized page design in e-commerce and stimulate further investigation in this area.