Critical node detection in temporal social networks, based on global and semi-local centrality measures

Area of Science:

• Network Science
• Epidemiology
• Data Analysis

Background:

• Critical nodes are vital for network function, influencing spread dynamics.
• Identifying these nodes is crucial in temporal networks, where connections change over time.
• Existing methods struggle to accurately pinpoint critical nodes in dynamic network structures.

Purpose of the Study:

• To develop and evaluate novel measures for critical node detection in temporal networks.
• To assess the effectiveness of these new measures in controlling epidemic spread and information dissemination.
• To compare the performance of proposed methods against established centrality measures.

Main Methods:

• Proposed three new metrics: temporal supracycle ratio, temporal semi-local integration, and temporal semi-local centrality.
• Analyzed measure performance using the SIR (Susceptible-Infected-Recovered) epidemic model.
• Compared results with existing temporal centrality measures like betweenness, centrality, and degree deviation.

Main Results:

• The proposed measures accurately identify influential nodes in temporal networks.
• Effectively demonstrated utility in SIR model scenarios: isolation, seeding, and network attack.
• Outperformed existing temporal centrality measures in identifying critical nodes.

Conclusions:

• The novel measures provide a more accurate way to detect critical nodes in temporal networks.
• These methods can enhance epidemic control strategies by identifying nodes for isolation.
• The proposed methods can also optimize information dissemination by selecting effective initial spreaders.