Best influential spreaders identification using network global structural properties.

Area of Science:

• Complex network analysis
• Information and epidemiological spread modeling

Background:

• Influential spreaders are critical nodes in complex networks for controlling or maximizing spread.
• Existing methods for identifying influential spreaders struggle with diverse network connectivity structures (complete, incomplete, in-between).
• No single indexing strategy is universally sufficient for all network types.

Purpose of the Study:

• To propose a novel indexing method, Network Global Structure-based Centrality (ngsc), for identifying influential spreaders.
• To enhance the accuracy and performance of spreader identification across various network connectivity structures.
• To provide a more robust method that accounts for global network properties.

Main Methods:

• Developed the Network Global Structure-based Centrality (ngsc) method.
• Integrated existing k-shell and sum of neighbors' degree methods.
• Incorporated global structural properties: giant component, average degree, and percolation threshold.
• Evaluated performance using an SIR (Susceptible-Infected-Recovered) model as ground truth.

Main Results:

• The ngsc method demonstrates superior spreading performance of seed spreaders across a wide range of network structures.
• Results show good correlation with rankings derived from the SIR model.
• The proposed method outperforms contemporary techniques and is competitive with computationally expensive advanced approaches.

Conclusions:

• The ngsc method offers improved accuracy in identifying influential spreaders compared to existing techniques.
• This approach effectively handles diverse network connectivity structures by considering global properties.
• ngsc provides a computationally efficient and effective solution for influential spreader identification in complex networks.