A centrality measure for cycles and subgraphs II

Area of Science:

• Network Science
• Graph Theory
• Mathematical Physics

Background:

• Existing network analysis methods often focus on individual nodes.
• Quantifying the importance of entire groups of nodes within complex networks remains a challenge.
• Eigenvector centrality is a key metric for node importance but is vertex-specific.

Purpose of the Study:

• To introduce a novel group centrality measure for complex networks.
• To provide rigorous mathematical foundations for the proposed group centrality.
• To demonstrate the efficacy of the group centrality measure in real-world network analysis.

Main Methods:

• Development of a semi-commutative extension of a number theoretic sieve.
• Mathematical derivation of group centrality from network flow interception.
• Comparison with existing group centrality measures (Everett and Borgatti).

Main Results:

• The proposed group centrality measure ranges from 0 to 1, representing the fraction of network flows intercepted by a group.
• The group centrality measure is shown to induce eigenvector centrality on individual vertices.
• The new centrality measure successfully distinguishes protein complexes in the yeast interactome network.

Conclusions:

• The developed group centrality offers a robust extension of eigenvector centrality to node groups.
• This measure provides a powerful tool for analyzing the collective importance of nodes in complex systems.
• The application to biological networks highlights its potential in identifying functional modules like protein complexes.