HAP-SAP: Semantic Annotation in LBSNs using Latent Spatio-Temporal Hawkes Process

Abstract

e prevalence of location-based social networks (LBSNs) has eased the understanding of human mobility pa erns. Knowledge of human dynamics can aid in various ways like urban planning, managing tra c congestion, personalized recommendation etc. ese dynamics are in uenced by factors like social impact, periodicity in mobility, spatial proximity, in uence among users and semantic categories etc., which makes location modelling a critical task. However, categories which act as semantic characterization of the location, might be missing for some check-ins and can adversely a ect modelling the mobility dynamics of users. At the same time, mobility pa erns provide a cue on the missing semantic category. In this paper, we simultaneously address the problem of semantic annotation of locations and location adoption dynamics of users. We propose our model HAP-SAP, a latent spatio-temporal multivariate Hawkes process, which considers latent semantic category in uences, and temporal and spatial mobility pa erns of users. e model parameters and latent semantic categories are inferred using expectation-maximization algorithm, which uses Gibbs sampling to obtain posterior distribution over latent semantic categories. e inferred semantic categories can supplement our model on predicting the next check-in events by users. Our experiments on real datasets demonstrate the e ectiveness of the proposed model for the semantic annotation and location adoption modelling tasks.