Research And Implementation Of The Important Target User Mining System Based On The Telecommunication Communication Behavior Network

Telecommunication network is an important base station to realize long-distance communication,which plays an indispensable role in people’s life.However,in recent years,with the frequent occurrence of criminal activities in telecommunication network,the supervision and governance of the network has become increasingly critical.Because of the exponential growth of data in the telecommunication network and the variety of communication means,the regulation and governance means based on communicative content are increasingly weak.The communicative behavior of telecom network users can be considered as a kind of behavioral information.Compared with traditional content-based supervision methods,information mining based on the communicative behavior of telecom network users greatly reduces the dimension of information needed to be analyzed,and the important target refers to the users or user groups with important value in the network.These important users or groups can have a huge impact on the structure and functionality of the network.Data mining based on the communicative behavior data of telecom network users can provide a more macro perspective for the supervision and governance of telecom network by relevant departments.Therefore,based on the practical application requirements of the important target mining system of the communicative behavior of telecom network users,this paper carries out three studies,the main contents of which are as follows:(1)Propose a node importance ranking algorithm combining local information and global information of nodes.Aiming at the problem that the current node sorting algorithm can not meet the system’s requirements of mining important users in terms of sorting accuracy and time complexity,a node importance sorting algorithm based on iterative K-shell and improved information entropy was proposed.Firstly,the global information of nodes in the network is obtained by analyzing the iterative information in k-shell decomposition process.Secondly,an improved node information entropy is proposed to obtain the local information of nodes.Finally,the importance of nodes is ranked by combining the global and local information of nodes.By comparing the proposed algorithm with six other algorithms on four real data sets,it is found that the proposed algorithm can be better applied to large-scale networks with higher resolution,stronger node information dissemination ability and lower time complexity than the existing methods.(2)Propose a community discovery algorithm to improve the idea of tag propagation.In order to solve the problem that the current community discovery algorithm cannot meet the requirements of large-scale network community discovery in terms of the stability and time complexity of partition results,an improved label propagation algorithm NILPA based on modular gain and node information was proposed by combining modular gain and node information with LPA algorithm.Firstly,NILPA algorithm overcomes the randomness of LPA initial label allocation by maximizing modularity gain.Secondly,NILPA proposed an improved node information measure based on node degree and node importance ranking obtained by K-shell algorithm to overcome the randomness of LPA in label selection.NILPA algorithm is compared with other three algorithms on two real world networks and three artificial generated networks.The results show that NILPA algorithm has the best effect in large networks and the community partition result is stable,which can be better applied to large networks.(3)Design and implement the important target mining system based on the communicative behavior of telecom network users.The system based on telecommunication network user’s guide information for data mining,can guide bearing huge amounts of data storage,pretreatment and analysis tasks,at the same time system by using the research content and the proposed algorithm in data mining,to dig out the potential of key users and important groups through visual function to display.The system test results show that the system can provide important data support for the administration and supervision of telecommunication network.