Open AccessThe generalized centrality method for analyzing network cyberspace
Author(s) -
Горлушкина Наталия Николаевна,
Иванов Сергей Евгеньевич,
Иванова Любовь Николаевна
Publication year - 2019
Publication title -
kibernetika i programmirovanie
Language(s) - English
Resource type - Journals
ISSN - 2644-5522
DOI - 10.25136/2306-4196.2019.2.23117
Subject(s) - centrality , cyberspace , popularity , computer science , network analysis , graph , katz centrality , network science , theoretical computer science , social network analysis , betweenness centrality , data mining , artificial intelligence , information retrieval , the internet , complex network , mathematics , world wide web , psychology , statistics , social psychology , physics , quantum mechanics , social media
The subject of the research is the methods of network cyberspace analysis based on graph models. The analysis allows to find leaders of groups and communities, to find cohesive groups and visualize the results. The main methods of the graph theory used for cyberspace social networks are the methods of analyzing the centrality to determine the relative weight or importance of the vertices of the graph. There are known methods for analyzing centralities: by degree, by proximity, by mediation, by radiality, by eccentricity, by status, eigenvector, referential ranking. The disadvantage of these methods is that they are based only on one or several properties of the network participant. Based on the centrality analysis methods, a new generalized centrality method is proposed, taking into account such participant properties as the participant's popularity, the importance and speed of information dissemination in the cyberspace network. A mathematical model of a new method of generalized centrality has been developed. Comparison of the results of the presented method with the methods of the analysis of centralities is performed. As a visual example, a subgroup of cyberspace consisting of twenty participants, represented by a graph model, is analyzed. Comparative analysis showed the accuracy of the method of generalized centrality, taking into account at once a number of factors and properties of the network participant.