CARec: Content-Aware Point-of-Interest Recommendation via Adaptive Bayesian Personalized Ranking

Abstract

Location-based social networks (LBSNs) offer researchers user-generated content data to study users’ intrinsic patterns of preference. One important application of such study is to provide a personalize point-of-interest (POI) recommender system to improve users’ experience in LBSNs. However, most of the existing methods provide limited improvements on POI recommendation because they separately employ textual sentiment or latent topic and ignore the mutual effect between them. In this paper, we propose a novel content-aware POI recommendation framework via an adaptive Bayesian Personalized Ranking. First, we make full use of users’ check-in records and reviews to capture users’ intrinsic preference (i.e., check-in, sentiment, and topic preferences). Then, by aggregating users’ intrinsic preferences, we devise an adaptive Bayesian Personalized Ranking to generate the personalized ranked list of POIs for users. Finally, extensive experiments on two real-world datasets demonstrate that our framework significantly outperforms other state-of-the-art POI recommendation models in various metrics.