Accuracy and Precision of Methods for Community Identification in Weighted Networks

Abstract

Based on brief review of approaches for community identification and measurement for sensitivity characterization, the accuracy and precision of several approaches for detecting communities in weighted networks are investigated. In weighted networks, the community structure should take both links and link weights into account and the partition of networks should be evaluated by weighted modularity $Q^w$. The results reveal that link weight has important effects on communities especially in dense networks. Potts model and Weighted Extremal Optimization (WEO) algorithm work well on weighted networks. Then Potts model and WEO algorithms are used to detect communities in Rhesus monkey network. The results gives nice understanding for real community structure.