Analysis And Implementation Of User Relationship Based On Mobile CDR

Social network analysis recently has been getting more and more attention of scholars from various fields. Studying the influence maximization problem of the mobile social network, which belongs to the category of social network, has important significance for precise marketing, network optimization, public security and so on. This paper is based on the CDR(calling detail records) data provided by the signaling monitoring platform in the mobile communication network, and analyzes the user relationship of mobile social network on user level and cell level, and applies the result to top-k nodes mining.On user level, combined with the characteristics of the mobile communication network and based on the Hybrid Potential-influence Greedy Algorithm framework, this paper puts forward an improved information diffusion model and implements the top-k users mining, with modeling the user relationship tightness and user threshold and redefining the calculation of the user influence and the potential impact, and optimizing the process of the greedy phase,which improves the operation efficiency of the algorithm. On cell level, according to the communication relationship among cells, this paper puts forward a novel flood-like information diffusion model based on the flooding method of broadcasting in data network, and demonstrates the independency of the order and the convergence in information diffusion process, and proposes a flood-like greedy algorithm to realize the implementation of mining top-k cells, combining the Hill Climbing Greedy Algorithm.Based on the CDR data provided by the signaling monitoring platform and the mining algorithms studied in this paper, this paper uses Flex and Java technology to set up a user mining platform, and applies the research results of top-k users mining and top-k cells mining into the platform, and provides good effect of network topology display and reasonable interaction functions.