Node Importance Identification for Temporal Networks Based on Optimized Supra-Adjacency Matrix

Area of Science:

• Network Science
• Complex Systems Analysis

Background:

• Node importance identification is crucial for understanding temporal network dynamics.
• Existing methods like SAM and SSAM have limitations in capturing complex temporal network structures.

Purpose of the Study:

• To propose an optimized supra-adjacency matrix (OSAM) modeling method for temporal networks.
• To accurately express temporal network structure and node importance considering intra- and inter-layer relationships.

Main Methods:

• Developed an optimized supra-adjacency matrix (OSAM) incorporating edge weights for intra-layer relationships.
• Modeled directional inter-layer relationships using directed graph characteristics.
• Calculated node importance using an index based on eigenvector centrality across layers.

Main Results:

• The OSAM method accurately represents temporal network structures.
• OSAM demonstrated faster message propagation and larger message coverage compared to SAM and SSAM.
• Achieved better SIR (Susceptible-Infected-Recovered) and NDCG@10 (Normalized Discounted Cumulative Gain) indicators.

Conclusions:

• The OSAM method provides a more effective approach for node importance identification in temporal networks.
• OSAM's ability to consider weighted and directional relationships enhances its performance in dynamic network analysis.