Mining individual mobility patterns from mobile phone data

Abstract

Data collected from mobile phones have the potential knowledge to provide with important behavior patterns of individuals. In the current research, there is relatively few commercial software or application systems to fully meet the requirements of effectively mining these mobility patterns. In this paper, we firstly introduce two basic mobility concepts of stop and trajectory to support mobile phone data mining applications. We then present framework of individual mobility pattern mining based on mobile phone location information. Some approach and algorithms are also proposed to discover four basic types of pattern, such as stay pattern, frequent pattern, association pattern, and movement prediction. We use real mobile phone data to perform functions for discovering individual behavior patterns and demonstrate effectiveness of our framework.