讲座:Generative AI for Modeling Customer Behavior: Exploring Effective Intervention Strategies 发布时间:2024-11-22

 宾:韩雪雯 PhD. Candidate 清华大学

主持人:刘佳璐 助理教授 上海交通大学安泰经济与管理学院

 间:20241129日(周13:30-15:00

 点:上海交通大学徐汇校区安泰楼A511

 

内容简介:

This study explores the application of generative AI (GenAI) models for intervention analysis based on customer growth sequences in retail banks. Developing effective intervention strategies poses significant challenges due to the vast volume of customer behaviors and diverse interventions. Accurate predictions of future customer behaviors and identification of optimal intervention strategies are imperative for marketers. Traditional methods, such as recurrent neural networks often fail to capture long sequences and predict future behaviors, especially after multiple interventions. To address these issues, this study introduces a transformer-based model, designed to model long and complex customer behavior sequences. By integrating domain knowledge with tokenization techniques to construct financial events, Our model can analyze complex behaviors in a unified framework. Collaborating with a retail bank in China, we validate the models effectiveness through both online and offline experiments aimed at enhancing customer assets under management. Techniques such as the masked language modeling for intervention analysis provide deeper insights into various strategiesimpacts in heterogeneous conditions. This study significantly improves traditional methods and contributes to both academic literature and practical applications by leveraging generative AI to model sequential data, offering a new research paradigm for marketing interventions.

 

演讲人简介

Xuewen Han is a Ph.D. candidate in Department of Management Science and Engineering at Tsinghua University. She earned dual bachelors degrees in management and economics from Jilin University in 2019. Xuewens research centers on digital innovation, with a particular emphasis on fintech, where she leverages advanced methods like generative AI to address real-world business challenges through field experiments. Her working papers are currently under revision at top-tier information systems journals. Xuewen has also presented and published her work at leading conferences in information systems and computer science, including ICIS, INFORMS, CIST, PACIS, CNAIS, and ACM ICAIF.

 

欢迎广大师生参加!