Publication:
New Betweenness Centrality Node Attack Strategies for Real-World Complex Weighted Networks
New Betweenness Centrality Node Attack Strategies for Real-World Complex Weighted Networks
No Thumbnail Available
Files
Date
2021
Authors
Quang Nguyen
Ngoc-Kim-Khanh Nguyen
Davide Cassi
Michele Bellingeri
Journal Title
Journal ISSN
Volume Title
Publisher
Research Projects
Organizational Units
Journal Issue
Abstract
In this work, we introduce a new node attack strategy removing nodes with the highest conditional weighted betweenness centrality (CondWBet), which combines the weighted structure of the network and the node’s conditional betweenness. We compare its efficacy with well-known attack strategies from literature over five real-world complex weighted networks. We use the network weighted efficiency (WEFF) like a measure encompassing the weighted structure of the network, in addition to the commonly used binary-topological measure, i.e., the largest connected cluster (LCC). We find that if the measure is WEFF, the CondWBet strategy is the best to decrease WEFF in 3 out of 5 cases. Further, CondWBet is the most effective strategy to reduce WEFF at the beginning of the removal process, whereas the Strength that removes nodes with the highest sum of the link weights first shows the highest efficacy in the final phase of the removal process when the network is broken into many small clusters. These last outcomes would suggest that a better attacking in weighted networks strategy could be a combination of the CondWBet and Strength strategies.
Description
Keywords
CondWBet,
WEFF,
Network Robustness Measures,
Attack Strategies with WEFF