Mining frequency pattern from mobile users

Abstract

Group pattern was introduced to find groups of mobile users associated by means of physical distance and amount of time spent together. This paper addresses the inherent problem of group pattern, that mobile user are often not physically close together when they use mobile technology, by proposing frequency pattern. Frequency pattern use creative method to calculate frequency of communication between mobile users. By using frequency rather than physical distance, the closeness of two mobile users can better be represented. Performance of the proposed method indicates a suitable segment size and alpha value needs to be selected to get the best result.