Community Detection is one of the most investigated problems as it finds application in many real-life areas. However, detecting communities and analysing community structure are very computationally expensive tasks, especially on large networks. In light of this, to better manage large networks, two new Multi-Level models are proposed in order to reduced and simplify the original graph via aggregation of groups of nodes. Both models have been applied on two variants of an immune-inspired algorithm, the first one based on a fully random-search process, and the second based on a hybrid approach. From the experimental analysis it clearly appears that the two proposed models help the random-search and the hybrid immune-inspired algorithms to significantly improve their performances from both computational and quality of found solutions point of view. In particular, the hybrid variant appears to be very competitive and efficient.

A Comparative Analysis of Different Multilevel Approaches for Community Detection

Scollo R. A.;Cutello V.;Pavone M.
2023-01-01

Abstract

Community Detection is one of the most investigated problems as it finds application in many real-life areas. However, detecting communities and analysing community structure are very computationally expensive tasks, especially on large networks. In light of this, to better manage large networks, two new Multi-Level models are proposed in order to reduced and simplify the original graph via aggregation of groups of nodes. Both models have been applied on two variants of an immune-inspired algorithm, the first one based on a fully random-search process, and the second based on a hybrid approach. From the experimental analysis it clearly appears that the two proposed models help the random-search and the hybrid immune-inspired algorithms to significantly improve their performances from both computational and quality of found solutions point of view. In particular, the hybrid variant appears to be very competitive and efficient.
2023
978-3-031-26503-7
978-3-031-26504-4
Community detection
Hybrid metaheuristics
Immune-inspired computation
Metaheuristics
Multi-level search
Random search
File in questo prodotto:
Non ci sono file associati a questo prodotto.

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/20.500.11769/553123
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact