A Path-Based Model for Link Prediction in Location-Based Social Networks

Abstract

In recent years, link prediction is considered as one of the prominent task in studying the social networks. Link prediction problem also has applicability in other domains like health care, information retrieval, e-commerce, and bioinformatics. With the development and comprehensive utilization of the compact devices, location-based social network (LBSN) has grown a critical stand for several novel applications since location data will help to derive many potential relationships in various domains. Friendship prognostication in standard interfaces is beneficial for many applications, such as friend or place recommendation and security administration. For loss of standard friendship prognostication algorithms, in this paper, three new methods were proposed for friendship prognostication by considering social and mobility patterns of users in LBSNs. Studies conducted on real-world datasets demonstrate that our propositions obtain a contentious performance with schemes from the research and, in a most maximum of the circumstances, exceed them.