LLM-Enhanced Reranking for Complementary Product Recommendation

2025-07-22

AI Summary

Complementary item recommendation faces a fundamental trade-off between accuracy and diversity, particularly underperforming for long-tail items. To address this, we propose a large language model (LLM)-enhanced re-ranking framework that, for the first time, directly leverages pre-trained LLMs—without fine-tuning or retraining—to re-rank candidate lists generated by graph neural networks (GNNs). Through carefully engineered prompts, the LLM captures user intent and collaborative semantic relationships among items, enabling joint optimization of precision and diversity. Extensive experiments on multiple public e-commerce datasets demonstrate that our method achieves over 50% average improvement in accuracy and up to 2% gain in diversity compared to state-of-the-art baselines, with notably enhanced long-tail item coverage. Our core contribution is the introduction of an LLM-driven lightweight re-ranking paradigm that balances computational efficiency, interpretability, and cross-domain generalizability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Recommender Systems
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Complementary product recommendation, which aims to suggest items that are used together to enhance customer value, is a crucial yet challenging task in e-commerce. While existing graph neural network (GNN) approaches have made significant progress in capturing complex product relationships, they often struggle with the accuracy-diversity tradeoff, particularly for long-tail items. This paper introduces a model-agnostic approach that leverages Large Language Models (LLMs) to enhance the reranking of complementary product recommendations. Unlike previous works that use LLMs primarily for data preprocessing and graph augmentation, our method applies LLM-based prompting strategies directly to rerank candidate items retrieved from existing recommendation models, eliminating the need for model retraining. Through extensive experiments on public datasets, we demonstrate that our approach effectively balances accuracy and diversity in complementary product recommendations, with at least 50% lift in accuracy metrics and 2% lift in diversity metrics on average for the top recommended items across datasets.

Problem

• Improving accuracy-diversity tradeoff in complementary product recommendations
• Leveraging LLMs for reranking without model retraining
• Enhancing long-tail item performance in e-commerce suggestions

Innovation

• LLM-enhanced reranking for complementary products
• Model-agnostic LLM prompting for reranking
• Balances accuracy and diversity without retraining