Effectiveness of LLMs in Temporal User Profiling for Recommendation

2025-10-31

AI Summary

This paper addresses the challenge of dynamically modeling user preferences in recommender systems. We propose a dual-temporal user profiling method leveraging large language models (LLMs): user interaction histories are first converted into textual summaries, then processed via an attention mechanism that explicitly disentangles short-term interests from long-term preferences, yielding natural-language-based, discriminative user profiles. The approach simultaneously enhances recommendation accuracy and ensures intrinsic interpretability. Experiments demonstrate strong performance in domains with dense user interactions and pronounced temporal patterns—e.g., movie and TV show recommendation—while exhibiting limited gains in sparse-interaction settings such as gaming, revealing a fundamental trade-off between temporal modeling efficacy, data density, and computational cost. Our core contribution is the first integration of LLM-driven natural-language summarization with fine-grained temporal preference disentanglement, unifying performance improvement and interpretability within a single framework.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Effectively modeling the dynamic nature of user preferences is crucial for enhancing recommendation accuracy and fostering transparency in recommender systems. Traditional user profiling often overlooks the distinction between transitory short-term interests and stable long-term preferences. This paper examines the capability of leveraging Large Language Models (LLMs) to capture these temporal dynamics, generating richer user representations through distinct short-term and long-term textual summaries of interaction histories. Our observations suggest that while LLMs tend to improve recommendation quality in domains with more active user engagement, their benefits appear less pronounced in sparser environments. This disparity likely stems from the varying distinguishability of short-term and long-term preferences across domains; the approach shows greater utility where these temporal interests are more clearly separable (e.g., Movies&TV) compared to domains with more stable user profiles (e.g., Video Games). This highlights a critical trade-off between enhanced performance and computational costs, suggesting context-dependent LLM application. Beyond predictive capability, this LLM-driven approach inherently provides an intrinsic potential for interpretability through its natural language profiles and attention weights. This work contributes insights into the practical capability and inherent interpretability of LLM-driven temporal user profiling, outlining new research directions for developing adaptive and transparent recommender systems.

Problem

• Modeling dynamic user preferences for improved recommendation accuracy and transparency
• Distinguishing short-term interests from long-term preferences in user profiling
• Evaluating LLM effectiveness across domains with varying user engagement patterns

Innovation

• LLMs generate short-term and long-term user summaries
• LLMs capture temporal dynamics in user preferences
• LLM profiles provide interpretability through natural language