关键词: 个性化推荐, 基于位置的服务, 协同过滤, 社交网络, 关联推荐

Abstract: Photos in Location-Based Social Networks（LBSN） include rich time and space location information, which provides conditions for exploring user preference information and attraction recommendations. The existing recommendation methods have many problems such as single recommended conditions, difficulty on estimation of user preference, low accuracy of recommendations. For those problems, PTLR method is proposed to improve the traditional method. Experimental results on LBSN real data show PTLR can effectively improve the recommendation accuracy.

Key words: personalized recommendation, location-based service, collaborative filtering, social networks, association recommendation