Open AccessAn Empirical Study of Human Mobility Patterns
Author(s) -
Douglas do Couto Teixeira,
Jussara M. Almeida
Publication year - 2018
Language(s) - English
Resource type - Conference proceedings
DOI - 10.5753/sbrc.2018.2411
Subject(s) - computer science , mobility model , work (physics) , data mining , empirical research , data science , machine learning , artificial intelligence , information retrieval , computer network , statistics , mechanical engineering , mathematics , engineering
This paper documents our efforts towards understanding which factors are more relevant in human mobility prediction. Our work is divided into two phases. First, we characterize a dataset consisting of more than 200,000 user check-ins in the Foursquare social network, inferring important patterns in human mobility. Second, we use factorial design to quantify the importance of several types of contextual information in human mobility prediction. Our results show that the proximity of the users possible next check-in to his or her home and work location are the most important factors (among the ones we analyzed) to be used by mobility prediction models.