Zero-Shot Next-Item Recommendation using Large Language Models

Anonymous

Abstract: Large language models (LLMs) have demonstrated impressive performance in various natural language processing tasks when given appropriate input prompts without requiring fine-tuning on specific training data. However, their application in next-item recommendation remains unexplored due to the vast, task-specific recommendation space and unfamiliarity with user preferences. To address these issues, this paper introduces the \textbf{Zero-Shot Next-Item Recommendation (NIR)} strategy, using an external module for candidate item generation and a \textit{3-step prompting} method for capturing user preferences and making ranked recommendations. Evaluations on MovieLens 100K and LastFM datasets using GPT-3.5 reveal that the proposed NIR competes well with strong sequential recommendation models, opening up new interesting research opportunities to leverage LLMs as recommender systems.

Paper Type: short

Research Area: NLP Applications

Contribution Types: Model analysis & interpretability, NLP engineering experiment

Languages Studied: English

Consent To Share Submission Details: On behalf of all authors, we agree to the terms above to share our submission details.