Fu Cai, Xiao-Yang Liu, Li Chen, and Qingfeng Huang
User behavior, movement pattern, data mining, similarity metric
Recent years, wireless networks are deeply integrated and embedded in our daily life and will be more personalized for diversified needs of data transmission and information sharing. A deep understanding of users’ behavior is bound to benefit the design, operation and maintenance of wireless communication networks. In addition, pervasive wireless networks initiate a new era of sociology by providing high granularity of user movement logging and complete trace sets, from which sociologists are interested in recognizing users’ mobility pattern. In this paper, a new method is proposed to analyze users’ behavioral patterns, which adopts time series to represent users’ movement, computes users’ similarity, and utilizes unsupervised clustering algorithm to find the data structure of potential features. First, we recognize certain social structures through analyzing log files of different users, rather than just performing statistic work. Second, we propose a similarity metric, which is suitable for mining users’ behavioral patterns in wireless networks. Third, we test our scheme on real data sets. Our scheme serves as first steps towards better network management, and behavior-aware network protocols, services and applications, etc.
Important Links:
Go Back
