Community Classification Based On Communication Data

In recent years, the classification of objects has been widely used in many research fields, and the classification method has also been greatly developed, such as clustering analysis, KNN algorithm, decision tree, support vector machine and so on.This paper focuses on the community classification based on communication data.We propose a complementary approach, which combines PageRank algorithm and SimRank algorithm, applicable in any domain with object-to-object relationships. We use this method in two practical cases, and compare the results with the real values and traditional clustering results respectively, our result and interpretation are ideal.Our method is applicable in any domain with object-to-object relationships.Firstly, we translate the problem into a simple and intuitive point-side structure graph model, then calculate the "importance" of status point by the PageRank algorithm and the "similarity" between any pair of status point's by the SimRank algorithm. Finally,the effective classification is obtained according to their relationship with each other.The basic idea of this paper is that "two objects are similar if they are related to similar objects."Our research work can be regarded as the practice and exploration of unsupervised learning.