Application Of Graph Algorithm Based On Hadoop In Telecom Social Network Analysis

Big data processing technology has emerged in the data era,with general speed improvement and development.In such a context,all walks of life also can not wait to introduce this technology to their own field.The telecommunications industry holds the pulse of the people's daily communication,with hundreds of millions of users each hour to produce a user calls,consumption,the Internet and other data.The application of these data to telecom operators can be analyzed from different perspectives,in order to support the data base of their business operations.One of the most important application is the analysis of the social network formed by their user.The use of big data processing technology can be very conducive to improve the speed of data processing which telecom operators own.Today,Hadoop MapReduce can be recognized for the best tool for big data processing.In this paper,we use the computational framework to parallelize the social network analysis algorithm.Based on the demand of determing the transmission value of some specified users in telecom networks,we implement MR-BC algorithm by combining the algorithm of fast betweeness centrality with MapReduce.In view of the problems such as high complexity of the data space and iteration,by using principles and strategies of MapReduce programming framework,the MR-BC algorithm is optimized.Community is another important research point in social network analysis.Telecom operators can understand the communication behavior of the users by simulating the relationship between users in real life.In this paper,we have studied the CPM which are used in different types of networks popularly.The MR-CPM algorithm is designed and implemented in distributed computing model.And the distributed design idea and algorithm flow are described in detail in this paper.