Resumen

DIAZ MATOS, Armando  y  GRASS BOADA, Darian Horacio. A new framework for centrality on multilayer networks. RCCI [online]. 2022, vol.16, n.3, pp. 1-17.  Epub 01-Sep-2022. ISSN 2227-1899.

The analysis of the relevance of the entities that make up a network is a task that allows to analyze, understand and even predict the future of the members of the network as well as the flow and scope of the information. Are called multilayer networks those that have several types of relationships between their entities, an example of which are social networks where users connect with different types of relationships, whether they are family, work or etc. Calculating the relevance or centrality of entities in multilayer networks is a challenging task due to the multiple ways of analyzing this concept and the computational cost involved in calculating it in large networks.

This paper presents a tool for calculating centrality measures in multilayer networks. It used Scala language and the Spark framework with the purpose of calculating in a parallel and distributed way the different centrality metrics. In the analysis of relevance, the different semantic levels determined by a selection of the types of edges as well as their merge are considered. The experiments speed up to 28.65 and 25.48 times for the measure of Closeness and Intermediation, respectively. Similarly, but to a lesser extent, the Eigenvector metrics reached accelerations of 6.91 times. The results show that the measures of centrality implemented scale.

Palabras clave : Multilayer Networks; Centrality Measures; Spark Framework.

        · resumen en Español     · texto en Español     · Español ( pdf )