Abstract
Because of the widespread adoption of GPS-enabled devices, such as smartphones and GPS navigation devices, more and more location information is being collected and available. Compared with traditional ones (e.g., Amazon, Taobao, and Dangdang), recommender systems built on location-based social networks (LB-SNs) have received much attention. The former mine users' preferences through the relationship between users and items, e.g., online commodity, movies and music. The latter add location information as a new dimension to the former, hence resulting in a three-dimensional relationship among users, locations, and activities. In this article, we summarize LBSN recommender systems from the perspective of such a relationship. User, activity, and location are called objects, and recommender objectives are formed and achieved by mining and using such 3D relationships. From the perspective of the 3D relationship among these objects, we summarize the state-of-the-art of LBSN recommender systems to fulfill the related objectives. We finally indicate some future research directions in this area.