Identifying highly influential nodes in multilayer networks based on global propagation.

Area of Science:

• Network Science
• Information Propagation
• Computational Social Science

Background:

• Information diffusion in networks is crucial for understanding social dynamics.
• Selecting optimal propagation sources is vital for maximizing information spread.
• Existing methods often suffer from computational inefficiency and simulation errors.

Purpose of the Study:

• To develop an efficient algorithm for identifying optimal propagation sources.
• To analyze the distribution of propagation ranges in single and multi-layer networks.
• To enhance the ranking of node influence for maximizing information spread.

Main Methods:

• Utilizing percolation theory and the independent cascade model.
• Developing a linear-complexity iterative algorithm for global propagation probability.
• Extending the algorithm to multi-layer networks, including two-layer networks.
• Proposing a de-overlapping method for multi-propagation source optimization.

Main Results:

• The iterative algorithm eliminates random errors and reduces computation time.
• Propagation range in single-layer networks follows a bimodal distribution.
• Propagation range in two-layer networks exhibits a four-peak distribution.
• The method enables direct calculation of peak probabilities and node influence ranking.

Conclusions:

• The proposed iterative algorithm is efficient and accurate for selecting optimal propagation sources.
• The analysis of propagation range distributions provides new insights into network diffusion.
• The methods enhance the ability to maximize information spread and influence in complex networks.