Community Clustering Algorithm on Semantic Similarity in Complex Network

—This paper proposes applying a WordNet based semantic similarity clustering algorithm on the cluster analysis of complex network community. By utilizing the semantic hierarchy of WordNet, the proposed method defines the key concept sets and the concept feature values for the community and constructs the semantic relations between concepts of the community nodes, which expands the application of clustering algorithms from text documents to complex network communities. The cluster structures derived from the proposed algorithm are concordance with peoples' judgments on a specific area, which will lead to the solution of the clustering problems in the complex network of different areas. Compared with VSM and k-MEANS, the proposed algorithm discovers more reasonable results and shows its effectiveness.