Point-of-interest recommendation model considering strength of user relationship for location-based social networks

Point of interest (POI) recommendation systems have drawn the attention of researchers in multiple domains, particularly location-based social networks (LBSNs). However, owing to barriers in data collection and information classification, most existing systems lack adaptability for users with varied relationship circles, which leads to unsatisfactory recommendation results. In this study, a model that considers user relationship strength is provided based on a data-driven method for improving the POI recommendations. It defines the user relationship according to an analysis of the user’s check-in behavior. The user’s social links are then embedded in the spatiotemporal model for POI recommendations. The effectiveness of this dynamic recommendation model is demonstrated by comparing six state-of-art POI recommendation techniques on three real-world datasets. The experiment results found significant correlations between the user relationship strength and check-in locations, which improved the model performance. Conceptually, this study supports the hypothesis that user relationship traits help explain personal preferences in LBSN usage and places visited. This study provides an intelligent social network system that provides real-time, location-aware recommendations for retailers.