An approach to cluster mobile users based on time and location

Abstract

Obtaining user location trails has become much easier as the coming of mobile internet. However, how to accurately cluster mobile users based on their location trails has become a challenge. This paper presents an approach to clustering mobile users based on time and location. It extracts center position from user location trials according to duration and uses time factor and time-weighting to cluster based on DBSCAN algorithm. The implementation of building circle by shaking in a campus social platform based on Android is conducted to confirm the effectiveness of the approach in this paper.