The purpose of this paper is to clarify the correlations between amount of individual’s knowledge of a specific area and his/her visit pattern to point of interest (POI, interested places) located in the area.
This paper proposes a visit-frequency-based familiarity estimation method that estimates individuals’ knowledge of areas in a quantitative manner. Based on the familiarity degree, individuals’ visit logs to POIs are divided into a set of groups followed by analyzing the differences among the groups from various points of view, such as user preference, POI categories/popularity, visit time/date and subsequent visits.
Existence of statistically significant correlations between individuals’ familiarity to areas and their visit patterns is observed by our analysis using 1.4-million POI visit logs collected from a popular location-based social network (LBSN), Foursquare. There exist different skewness of the visit time and visited POI distribution/popularity with regard to the familiarity. For instance, users go to unfamiliar areas on weekends and visit POIs for cultural experiences, such as museums. A notable point is that the correlations can be detected even in the areas in home city, which have not been known so far.
This is the first in-depth work that studies both estimation of individuals’ familiarity and correlations between the familiarity and individuals’ mobility patterns by analyzing massive LBSN data. The methodologies used and the findings of this work can be applicable not only to human mobility analysis for sociology, but also to POI recommendation system design.
Han, J. and Yamana, H. (2016), "A study on individual mobility patterns based on individuals’ familiarity to visited areas", International Journal of Pervasive Computing and Communications, Vol. 12 No. 1, pp. 23-48. https://doi.org/10.1108/IJPCC-01-2016-0010Download as .RIS
Emerald Group Publishing Limited
Copyright © 2016, Emerald Group Publishing Limited