Analysis of mobile subscribers' behavior pattern based on non-negative matrix factorization

As the increasing number of intelligent mobile phones that are available for people which lead to tones of network flow data are produced. Hence, how to dig the user's behavior pattern by using the data collected form mobile phones is a significant problem. While some achievements have been made in this area, it is still too complicated to be put into practice due to various conditions needed by traditional research method. Meanwhile, users' data of mobile communication network gradually improved, so that it only requires part of information to reach the ideal status. Hence, we propose a Non-negative Matrix Factorization based method to analyze the mobility patterns of mobile phone users, i.e., find the groups of people with the similar mobility patterns. The experimental results on the real data obtained from the capital of a south province validate our proposed method.