Real-world complex systems can be modeled as homogeneous or heterogeneous graphs composed by nodes connected by edges. The importance of nodes and edges is formally described by a set of measures called centralities which are typically studied for graphs of small size. The proliferation of digital collection of data has led to huge graphs with billions of nodes and edges. For this reason, we focus on two new algorithms, Game of Thieves and WERW-Kpath which are computationally-light alternatives to the canonical centrality measures such as degree, node and edge betweenness, closeness and clustering. We explore the correlation among these measures using the Spearman’s correlation coefficient on real criminal networks extracted from judicial documents of three Mafia operations. Results of our analysis indicate that Game of Thieves could be used as a more economic replacement to rank both nodes and edges and WERW-Kpath to rank edges.

Game of Thieves and WERW-Kpath: Two Novel Measures of Node and Edge Centrality for Mafia Networks

Ficara A.;Saitta R.;Fiumara G.;De Meo P.;Liotta A.
2021-01-01

Abstract

Real-world complex systems can be modeled as homogeneous or heterogeneous graphs composed by nodes connected by edges. The importance of nodes and edges is formally described by a set of measures called centralities which are typically studied for graphs of small size. The proliferation of digital collection of data has led to huge graphs with billions of nodes and edges. For this reason, we focus on two new algorithms, Game of Thieves and WERW-Kpath which are computationally-light alternatives to the canonical centrality measures such as degree, node and edge betweenness, closeness and clustering. We explore the correlation among these measures using the Spearman’s correlation coefficient on real criminal networks extracted from judicial documents of three Mafia operations. Results of our analysis indicate that Game of Thieves could be used as a more economic replacement to rank both nodes and edges and WERW-Kpath to rank edges.
2021
978-3-030-81853-1
978-3-030-81854-8
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11570/3218520
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? 4
social impact