Browsing Behavioral Intent Prediction on Product Recommendation Pages of E-commerce Platform

Abstract

User behavior data has always been the key for e-commerce platforms to make decisions and improve experience, especially when predicting users’ behavioral intent. Nowadays, the Product Recommendation Page (PRP) has played an increasingly significant role of e-commerce platforms with the popularity of recommendation systems. However, past research on predicting user behavioral intent across e-commerce platforms may not be applicable to PRPs, where users have different characteristics. In this research, users’ browsing behavioral intent of PRPs is studied and predicted. A large amount of user data of PRPs is collected and processed, and the corresponding dataset is built. After that, a user interest analysis method is proposed while five browsing intent prediction models are applied and compared. The method distinguishes users with different browsing interest degrees, and the models can better predict users’ browsing behavior intent within different interest groups. A validation experiment on the large-scale dataset shows that the proposed method can predict user browsing intent with a decent performance.