Patterns in Temporal Networks with Higher-Order Egocentric Structures

Area of Science:

• Complex Systems Science
• Network Science
• Social Network Analysis

Background:

• Temporal networks model time-evolving interactions but often focus on pairwise relationships.
• Existing egocentric temporal neighborhood measures neglect group interactions common in social systems.
• Higher-order networks offer a framework for analyzing group interactions.

Purpose of the Study:

• To generalize egocentric temporal neighborhood analysis to higher-order interactions using hypergraphs.
• To introduce the concept of "hyper egocentric temporal neighborhoods" for analyzing complex social dynamics.
• To assess the impact of higher-order interactions on temporal network analysis.

Main Methods:

• Generalizing temporal networks to hypergraphs to represent group interactions.
• Defining and applying "hyper egocentric temporal neighborhoods" for network decomposition.
• Analyzing temporal network data, including second-order interactions (triplets).

Main Results:

• The proposed hyper egocentric temporal neighborhoods effectively capture higher-order social interactions.
• Second-order structures (triplets) account for the majority of variability in the data.
• This variability is observed across different datasets, between nodes, and over time.

Conclusions:

• Higher-order representations are crucial for a comprehensive understanding of temporal social networks.
• Hyper egocentric temporal neighborhoods provide a powerful tool for analyzing complex, group-based interactions.
• The findings highlight the importance of moving beyond pairwise analysis in complex systems.