Persona guided cross domain recommendation improves user behavior predictions
When Semantics Fall Short: Persona-Guided Cross-Domain Sequential Recommendation with Large Language Models
Information Retrieval
Summary
Recommending products or content across different categories can be hard because what works in one area might not work in another. The paper finds that using global meanings of items alone isn’t enough to connect these different areas well. Instead, the authors suggest using imagined user profiles, called personas, created by large language models to better understand how people behave across categories. Their method combines these personas with local item relationships to make better recommendations. Tests show this approach works better than previous methods.
What this means in practice
- •For ecommerce platform teams: Improves cross-category product recommendations by using persona-guided behavior patterns to better predict user interests.$Commercial implications: Enables development of smarter recommendation engines for retailers that adapt across product categories, increasing sales and user engagement.
- •For media streaming services: Enhances user content suggestions across different media types by integrating persona-driven behavior transfer and local item relations.
Authors
Jundong Chen, Honglei Zhang, Shuai An, Enneng Yang, Han Yu
Abstract
Cross-domain sequential recommendation (CDSR) alleviates data sparsity by transferring user behaviors across domains. Recent advances increasingly leverage large language models (LLMs) to bridge domains through global item semantics. However, this paradigm is inherently limited by two issues: I. semantic domain isolation, where item semantics remain domain-specific and hinder transferable modeling, and II. semantic-behavior inconsistency, where local semantic relations are more consistent with behavioral transitions, whereas such consistency often breaks down globally across domains. To this end, we propose PerCDR, an LLM-enhanced CDSR framework that shifts the focus from global semantic bridging to persona-guided behavioral modeling. Specifically, PerCDR leverages LLM-inferred personas to globally guide transferable cross-domain behavioral patterns, while exploiting local semantic relations to enhance intra-domain behavioral modeling, followed by domain preference adaptation for recommendation refinement. Extensive experiments demonstrate the superiority of PerCDR and validate its effectiveness.