THE APPLICATION OF RANGE CLUSTERING METHOD IN COMMUNITY DETECTING PROBLEM

Abstract

Nowadays, community detection in graphs has been an important problem in computer science research. Social networks are often expressed in form of structure data graph. Hence, social network community mining mainly deals with graph clustering problem. To solve this problem, many algorithms have been proposed. In this article, the authors present new research results based on the approach of using the concept of spectrum to bring the problem of clustering general graph of clustering problem on vectors of real numbers only, for this reason, the number of dimensions of the data set will be reduced, then we incorporate the techniques of optimizing the Min-cut function using the Laplace matrix. This approach will reduce the calculation complexity and quickly yields the result of social network community structure mining. The effectiveness of proposed algorithm is evidenced by experimental results on real data sets.