Exploring the Effect of Context-Awareness and Popularity Calibration on Popularity Bias in POI Recommendations

2025-07-04

AI Summary

POI recommendation suffers from popularity bias, leading to the underrepresentation of niche yet meaningful locations. To address this, we propose a joint framework integrating context-aware modeling with popularity calibration. First, we empirically reveal the heterogeneous impact of diverse contextual factors—such as time, location, and social signals—on popularity bias. Building on this insight, we design a learnable popularity calibration module that dynamically aligns the recommendation distribution with users’ true preferences. Extensive experiments on four real-world POI datasets demonstrate that our approach significantly mitigates popularity bias while preserving recommendation accuracy, substantially improving coverage of long-tail POIs and enhancing personalized matching. Our key contributions are twofold: (i) the first systematic characterization of the context-popularity interaction mechanism, and (ii) the co-optimization of calibration intensity and context sensitivity within a unified learning framework.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences
• Data Mining & Knowledge Management: Recommender Systems

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Point-of-interest (POI) recommender systems help users discover relevant locations, but their effectiveness is often compromised by popularity bias, which disadvantages less popular, yet potentially meaningful places. This paper addresses this challenge by evaluating the effectiveness of context-aware models and calibrated popularity techniques as strategies for mitigating popularity bias. Using four real-world POI datasets (Brightkite, Foursquare, Gowalla, and Yelp), we analyze the individual and combined effects of these approaches on recommendation accuracy and popularity bias. Our results reveal that context-aware models cannot be considered a uniform solution, as the models studied exhibit divergent impacts on accuracy and bias. In contrast, calibration techniques can effectively align recommendation popularity with user preferences, provided there is a careful balance between accuracy and bias mitigation. Notably, the combination of calibration and context-awareness yields recommendations that balance accuracy and close alignment with the users' popularity profiles, i.e., popularity calibration.

Problem

• Evaluating context-aware models to mitigate POI popularity bias
• Assessing calibrated popularity techniques for bias reduction
• Combining context-awareness and calibration for balanced recommendations

Innovation

• Context-aware models for bias mitigation
• Calibration techniques aligning popularity
• Combining context-awareness and calibration