Search

The extraction of the communities is reflected by the allocation of the nodes to the different communities and therefore a distribution, to which we can associate a score indicating its quality. Several methods have been proposed to extract existing communities in a network. The classification of these methods may be the subject of a separate study. The common point noted between most of these methods is the use of a function that measures the quality of a distribution. A metric called "Modularity" is considered to be the most efficient; it is used in most of the community extraction algorithms. Other metrics have been proposed such as "Z-Modularity", "Modularity density", "Exponential modularity". However, Modularity is still a competing metric. Our job is to establish an in-depth study of these quality measurement functions. For this purpose, considering the case of non-overlapping communities, we thought of a structural comparison between the results obtained, executing the "Leiden" algorithm, each time we change the metric among the four adopted. We chose this algorithm because it is an improved version of one of the most popular algorithms in this area called "Louvain". [ABSTRACT FROM AUTHOR]