Telecommunication Relationship Based Opinion Leader Mining And Application

With the increasing competition in the telecommunication market home and abroad, the marketing method of telecommunication businesses gradually changes from extensive form to intensive mode. According to the characteristics of customer telecommunication data, this paper makes use of social network analysis techniques and data mining technologies, analyzing the features of opinion leader in telecommunication businesses customer groups and mining opinion leader in telecommunication network in order to meet the specific marketing needs of telecommunication businesses.First, we give a summary of the concepts, techniques and measurements of social network analysis. Analysis of telecommunication relationship network data shows that telecommunication relationship belongs to 1-mode type. Using social network analysis techniques, we discuss measurement features of opinion leader in telecommunication relationship network, which forms a theoretical basis for following discussing.Then, we analyze the features of telecommunication data and establish telecommunication relationship data warehouse according to the data warehouse theory. Based on the Six Degree Separation theory and the 150 rule, we construct the telecommunication relationship network. Computing models are compared and parameters of telecommunication relationship network are fixed. Using the Reciprocal Nearest Neighbor based algorithm, huge telecommunication relationship network is divided into several subnets. Extension degree, betweenness degree, closeness degree and centrality degree are computed and opinion leader nodes are mined out from subnets using the SPLINE algorithm based on the Dijkstra nearest path.Finally, in the mobile telecommunication business case of developing virtual network customers, we make use of opinion leader mining model to analyze the core customer. Application results show that our model outperforms in the aspects of the accuracy of mining core customers, the succeeding rate of developing virtual network, the quality of virtual network and covering rate of virtual network and etc. Application proves that it is feasible and effective to mine opinion leader in telecommunication relationship network using social network analysis techniques and data mining technologies.