Application of spectral methods for recognizing the structure of communities in complex networks

Abstract

In this paper, spectral clustering methods for detecting communities of an undirected graph are investigated. These algorithms are derived from graph partitioning problems and have become one of the most popular ways to determine structure in recent years. Several types of traditional spectral analysis algorithms have been implemented in Python programming language to identify communities in an undirected graph, and a comparative analysis of methods has been carried out, which will be unique information for the correct choice of a network structure detection method. The practical significance of the work lies in the possibility of the best choice of the implementation of the algorithm based on spectral methods for identifying communities, based on the properties of a particular network and the goals of partitioning.