Performance Evaluation of Community Detection Algorithms Based on Relationship Strength Measurement

The study and analysis of networks have been a significant field of research as it holds its roots in varied disciplines like Biology, Chemistry, Sociology, Computer Applications and many more. Human beings have ventured into the era of a network with the rise of all kinds of networks such as the Internet and Social networks. Detection of community structure in real networks is vital in terms of both theoretical and practical value. Many community detection methods are derived from specific backgrounds and their reliability is still questionable. Researchers hardly focus on the general definition of communities and the general community detection algorithms. The discrepancy between the two might pose some obstacles to the optimization of them so that it is hard to use one algorithm to perfect the other. Recently, Lu et al. [1] have proposed a general method for constructing network model from the real-world problem. Also, they have given a general definition of community structure as well as a complete procedure for detecting communities. In this paper, we apply the efficient resistance distance function [2] on Lu et al.'s algorithm and compare its performance with other community detection algorithms that allow for overlaps.