Analisis Perbandingan Kesentralan Graf Dengan Degree, Eigenvector, dan Beta Centrality

Abstract

The study of graph centrality in assessing the most central vertex in a graph or the most central or important individual in a network has been a prosperous field of exploration these days. Centrality observation in graphs is suitable for any graph with various kinds of centrality methods to be applied to many fields. In this study, simple, regular, directed and signed graph are being assessed by Degree, Eigenvector, and Beta Centrality. These three methods are related to each other; therefore, it is interesting to study the relation and the characteristic of each method. According to the result, Degree Centrality which assesses centrality based on vertexs direct links is applicable for every graph in this study. On the other hand, Eigenvector Centrality which asses a vertex centrality with respect to its neighbors centrality, is not applicable for the acyclic directed tree. While Beta Centrality is also advantageous for every graph in this study, the use of parameter β affects centrality scores depending on how the measure is conducted for local or global structure. Beta Centrality is an alternative to observing a vertexs centrality score by considering its direct links centrality, in the acyclic directed tree.