Edge Overlap in Weighted and Directed Social Networks

Area of Science:

• Social Network Analysis
• Computational Social Science
• Network Science

Background:

• Modern digital data sources (social media, cell phones) enable detailed analysis of social interactions.
• Traditional network metrics are often limited to unweighted and undirected networks, insufficient for rich, real-world data.
• Edge overlap, a key metric for shared connections, needs adaptation for complex network types.

Purpose of the Study:

• To extend the definition of edge overlap to weighted and directed social networks.
• To derive closed-form expressions for the mean and variance of these extended metrics.
• To provide a statistically sound framework for comparing edge overlap across diverse network structures.

Main Methods:

• Developed novel definitions for edge overlap in weighted and directed networks.
• Derived analytical expressions for mean and variance using Erdős-Rényi random graph models.
• Applied these methods to empirical social network data from rural India.

Main Results:

• Successfully extended edge overlap calculations to weighted and directed networks.
• Provided analytical formulas for mean and variance applicable to complex network models.
• Quantified deviations of empirical network overlap from random graph expectations.

Conclusions:

• Novel definitions enable edge overlap calculation for more complex network data.
• Derivations offer a statistically rigorous approach for cross-network comparisons.
• The study enhances the analytical toolkit for understanding social network structures.