A Partitioning Method for Community Structure in Weighted Networks Based on Node Similarity

Based on node similarity, a method for detecting community structure in a weighted network is proposed. A novel node similarity matrix of the weighted network is constructed, and then an arbitrary node is chosen as initial node based on it. A node having maximum similarity to the initial node is searched, and the two nodes are merged into a new community. The community structure is discovered iteratively, and the partition is formed eventually. The presented method has low computational complexity. Furthermore, the effectiveness of the algorithm is validated by the numerical examples of community detection with classic complex network.