HUMAN MOBILITY ANALYTICS WITH DEEP REPRESENTATIONS

Abstract

One of the largest dynamics of our society is the human mobility whose pricelessness is concealed by its mighty complexity and diversity, in both spatial and temporal aspects. A better understanding of such complex dynamics can benefit our livings in various perspectives. Benefit from big data, we are able to capture the complexity of human mobility using deep representation techniques. In this thesis, we conduct systematic analytics on human mobility by stratifying our view into three representation levels: location, trajectory, and aggregation level, each with corresponding analytic tasks to be addressed. At the location level, we unsupervisedly learn location embeddings with better semantic coherence by disentanglement. At the trajectory level, we capture and synthesize the geographic and semantic attributes completely for human traces. At the aggregation level, we consider citywide flow patterns and external factors for inferring fine-grained human flow data.