Research On Mobile User Group Identification Based On Data Mining Methods

With the rapid development of mobile communication network,people's social communication mode expands from the traditional social network to the mobile field,forming a mobile social network with instantaneous,convenient,mobile and personalized carrier for mobile devices.Since 2002,China mobile has shifted the focus of its customer service system to another aspect gradually,quietly and rapidly:group customers.Group customers have a strong strategic position in stabilizing high-value user groups,promoting mobile data and value-added services,and industrial applications.To gradually transform the competitive individual market into the competitive group customer market,operators need to provide more precise,differentiated and personalized services for the group users.Therefore,the accurate identification of group users is of great research value.In this paper,the actual data analysis of mobile communication was used to analyze the using regularity of mobile communication network,explore their calling habits,extract valuable character parameters according to the calling habit of the different groups of users,and mine the correlation of calling data between individual group and users.Finally,effective classification and discrimination methods were used to identify the identity of mobile users by experiment,and an accurate model for identity the group of each user.In this paper,the K-nearest neighbor method of the support vector machine method were used to analyze and compare.We found that the K-nearest neighbour method is more accurate and stable than the support vector machine method.