Community detection: Comparison of state of the art algorithms

Abstract

Real world complex networks may contain hidden structures called communities or groups. They are composed of nodes being tightly connected within those groups and weakly connected between them. Detecting communities has numerous applications in different sciences such as biology, social network analysis, economics and computer science. Since there is no universally accepted definition of community, it is a complicated task to distinguish community detection algorithms as each of them use a different approach, resulting in different outcomes. Thus large number of articles are devoted to investigating community detection algorithms, implementation on both real world and artificial data sets and development of evaluation measures. In this article several state of the art algorithms and evaluation measures are studied which are used in clustering and community detection literature. The main focus of this article is to survey recent work and evaluate community detection algorithms using stochastic block model.