Key Node Identification Method Based on Multilayer Neighbor Node Gravity and Information Entropy.

Area of Science:

• Complex network analysis
• Information propagation dynamics
• Network science

Background:

• Accurate identification of key nodes is vital for understanding and controlling information flow in complex networks.
• Existing local centrality methods may lack accuracy due to incomplete consideration of node-neighbor interactions.
• There is a need for robust methods to identify influential nodes in large-scale networks.

Purpose of the Study:

• To propose a novel method, multilayer neighbor node gravity and information entropy (MNNGE), for accurate key node identification.
• To enhance the analysis of information propagation by effectively capturing node interactions.
• To provide a parameter-free approach suitable for large-scale complex networks.

Main Methods:

• Calculating relative node gravity based on node weights.
• Computing direct node gravity by analyzing neighboring node attributes and local triangular structures.
• Aggregating relative and direct gravity using information entropy to derive node centrality.

Main Results:

• The MNNGE method demonstrated superior accuracy in identifying key nodes compared to existing methods across various real-world network datasets.
• Evaluation using metrics like the susceptible-infected-recovered (SIR) model and correlation coefficients confirmed MNNGE's effectiveness.
• The method requires no parameter settings, simplifying its application.

Conclusions:

• MNNGE offers a more accurate and efficient approach to key node identification in complex networks.
• The method's ability to consider multilayer neighbor interactions and its parameter-free nature make it highly applicable.
• MNNGE is well-suited for analyzing information propagation in large-scale complex systems.