Exploring the evolution, progress, and future of point-of-interest recommendation over location-based social network: a comprehensive review

Location-based social networks (LBSNs) have bridged the gap between the virtual and real worlds by allowing users to share their preferences and behaviors digitally. Point-of-Interest (POI) recommendation has become quite popular with the influx of geospatial data through LBSN. In this review, we discuss the growth and necessity of LBSN and trace back the evolution of POI recommendation systems. This review explicitly enunciates the factors impacting the recommendation process and the laws implicitly guiding the recommendation task. Unlike other reviews, we narrow down the categorization of the different models here based on subtle differences in the methodology, target user type, and target domain. We accentuate the representative works for each category based on their contributions, datasets, influencing factors, and performance metrics. Moreover, we discuss impending challenges, real-world implications, and perspectives on applying the POI recommendation. Finally, we conclude our survey by highlighting future opportunities in the recommendation task.