Improvement of Ego Network Detection Algorithm Based on Cluster Validity Index

摘要

With the application and popularity of social software, online social networks have become crucial for group security research. The social network community detection algorithm can divide the target group into different sub-communities according to the attributes and structure of the group members. Based on the ego network community detection algorithm, this paper adopts the method suitable for graph structure, and integrates the idea of cluster validity index with the community detection algorithm to complete the selection of the optimal number of communities. And compared three kinds of cluster validity indexes, this paper solves the problem that the number of communities needs to be artificially specified, and improves the accuracy of the community detection algorithm. Thus, targeted social network community detection becomes more practical.