TGTM: Temporal-geographical topic model for point-of-interest recommendation

Abstract

The wide spread use of location based social networks (LBSNs) and Micro-blogging services generated large volume of users’ check-in data, which consists of user ids, textual contents, posting timestamps, geographic information and so on. Point-of-interest (POI) recommendation is a task to provide personalized recommendations of interesting places to enhance the user experience in LBSNs. In this paper, we propose 2 novel time-location-content aware POI recommendation models which jointly integrate auxiliary temporal, textual and spatial information to improve the performance of POI recommendation. Specifically, we utilize temporal information to partition the original user-POI check-in frequency matrix into sub-matrices so that behavior in similar temporal scenario can be grouped. Then, we take advantage of Latent Dirichlet Allocation (LDA) model and spatial coordinates to infer the POIs. Comprehensive experiments conducted using real-world datasets demonstrate the superiority of our approach.