On community structure validation in real networks

Abstract Community structure is a commonly observed feature of real networks. The term refers to the presence in network groups nodes ( communities ) that high internal connectivity, but are poorly connected between each other. Whereas issue community detection has been addressed several works, problem validating partition as good for received considerably less attention and remains an open issue. We propose set indices validation partitions based on hypothesis testing procedure assesses distribution links within communities. Using both simulations data, we illustrate how proposed can be employed compare adequacy different structures given network, assess whether two networks share same or similar structures, evaluate performance clustering algorithms.