Estimating the Number of Communities Based on Maximum <i>a Posteriori</i>

Community structure is one of the important characteristics of complex networks. Community discovery or detection has important theoretical significance and practical value. At present, estimation of the number of communities (or clusters) is still an open question. By Bayesian inference, this research deduces the relationship between maximum possible partition and mutual information and information entropy, and proposes a framework algorithm Clustering Number Estimation (CNE) and a concrete implementation to estimate the number of clusters (K). Several typical algorithms for community number estimation are compared on several typical data sets, and primary experiments validate the effectiveness of this method.