[3I5-GS-11] AI and Society:

[3I5-GS-11-02] A Study on the TiSASRec Extension Model Capable of Considering Multiple User Behaviors and Customer Behavior Analysis

〇Tachi Kanie¹, Tenshou You², Shou Oiwa¹, Masayuki Goto¹ (1. Waseda University, 2. Keio University)

Keywords:sequential recommendation model, Interpretable AI, Customer behavior analysis, TiSASRec

In recent years, online platforms such as video streaming services have expanded rapidly. As a result, users are offered a vast number of choices, making it difficult for them to select the most suitable product or service for themselves.However, there is a risk that customer satisfaction with their choice will decline. Therefore, recommendation systems that recommend appropriate options to users are becoming increasingly important. In particular, sequential recommendation, which predicts and recommends items that the user is likely to purchase next based on the user's past behavior history, is attracting attention. One of the sequential recommendation methods is TiSASRec. This model can consider the time interval between each action in addition to the user's past action history. However, it is important to consider and incorporate behavior-based features other than the time interval between actions, such as content viewing time and number of clicks, in order to more accurately capture customer behavior and preferences. In this study, we propose a sequential recommendation model that can consider multiple user behaviors by extending TiSASRec. Furthermore, we conduct evaluation experiments and analysis using real data to demonstrate the effectiveness of the proposed method. We also discuss how to apply the proposed model to customer behavior analysis.