A bio-inspired methodology of identifying influential nodes in complex networks

Area of Science:

• Network Science
• Computational Biology
• Data Analysis

Background:

• Identifying influential nodes is crucial for understanding complex networks.
• Existing centrality measures like degree, betweenness, and closeness have limitations in capturing global network characteristics or are restricted to unweighted networks.
• Methods such as semi-local centrality, LeaderRank, and PageRank are not suitable for weighted networks.

Purpose of the Study:

• To propose a novel bio-inspired centrality measure for identifying influential nodes in weighted complex networks.
• To combine the strengths of Physarum centrality and K-shell decomposition for a more robust node influence assessment.
• To evaluate the proposed method's performance using the Susceptible-Infected (SI) model.

Main Methods:

• A bio-inspired centrality model integrating Physarum centrality with the K-shell index derived from K-shell decomposition analysis.
• Utilizing the Susceptible-Infected (SI) model for performance evaluation of the proposed centrality measure.
• Comparative analysis against existing centrality methods on example networks.

Main Results:

• The proposed model effectively identifies influential nodes in weighted networks.
• Performance evaluation using the SI model demonstrates the method's adaptivity and efficiency.
• The novel centrality measure shows improved results compared to traditional and existing advanced methods.

Conclusions:

• The integrated Physarum centrality and K-shell index model offers a superior approach for identifying influential nodes in weighted complex networks.
• The method provides a robust and efficient tool for network analysis with practical applications.
• This bio-inspired approach enhances the understanding of network structures and node importance.