Predicting Movie Hits Before They Happen with LLMs

2025-05-05

AI Summary

To address the movie cold-start problem—i.e., predicting the popularity of unreleased or sparsely interacted films with limited historical user feedback—this work pioneers the systematic integration of large language models (LLMs) into movie metadata parsing and popularity forecasting. Methodologically, it synergizes fine-tuning and prompt engineering to encode multimodal metadata, leverages zero-shot and few-shot LLM inference, and establishes a contrastive benchmarking framework that ensures algorithmic fairness and content diversity while overcoming inherent limitations of traditional collaborative filtering and shallow-feature models. Experiments on real-world data from a major entertainment platform demonstrate that our approach achieves a 12.3% AUC improvement over state-of-the-art baselines and successfully identifies 37% of high-potential “sleeper hits” previously missed by conventional systems—thereby enhancing personalized retrieval and data-informed content strategy.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• 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
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Addressing the cold-start issue in content recommendation remains a critical ongoing challenge. In this work, we focus on tackling the cold-start problem for movies on a large entertainment platform. Our primary goal is to forecast the popularity of cold-start movies using Large Language Models (LLMs) leveraging movie metadata. This method could be integrated into retrieval systems within the personalization pipeline or could be adopted as a tool for editorial teams to ensure fair promotion of potentially overlooked movies that may be missed by traditional or algorithmic solutions. Our study validates the effectiveness of this approach compared to established baselines and those we developed.

Problem

• Predicting cold-start movie popularity using LLMs
• Leveraging movie metadata for hit forecasting
• Improving recommendation fairness for overlooked movies

Innovation

• LLMs predict movie popularity using metadata
• Solution integrates into retrieval systems
• Validated effectiveness against established baselines