A split-and-transfer flow based entropic centrality

Area of Science:

• Network Science
• Information Theory
• Computational Social Science

Background:

• Entropic centrality measures node importance based on flow destination uncertainty.
• Existing models assume indivisible flow, limiting nuanced analysis.
• This limitation overlooks the potential for flow to split at network nodes.

Purpose of the Study:

• To propose a novel split-and-transfer flow model for entropic centrality.
• To enable arbitrary splitting of flow across neighboring nodes.
• To provide a computationally tractable method for analyzing this new centrality measure.

Main Methods:

• Developed a split-and-transfer flow model for entropic centrality.
• Mapped the new model to an equivalent transfer entropic centrality framework.
• Applied the model to empirical network datasets.

Main Results:

• The split-and-transfer model offers a more flexible and realistic approach to entropic centrality.
• The mapping to transfer entropic centrality simplifies computation.
• Case studies demonstrate the model's utility in diverse networks.

Conclusions:

• The split-and-transfer flow model advances entropic centrality by incorporating flow divisibility.
• This provides richer insights into node importance in complex networks.
• The method is applicable to various network types, including transportation, financial, and digital systems.