Social balance in directed networks

TL;DR

This paper extends traditional social balance theory to directed networks, introducing a maximum-entropy null model that preserves key features and reveals consistent balance patterns in large-scale signed directed social networks.

Contribution

It develops a maximally constrained null model for signed directed networks and demonstrates its effectiveness in uncovering social balance patterns.

Findings

Null model preserves directed topology and node degrees.

Balance patterns align with directed strong balance theories.

Consistent patterns found across large social networks.

Abstract

Social networks inherently exhibit complex relationships that can be positive or negative, as well as directional. Understanding balance in these networks is crucial for unraveling social dynamics, yet traditional theories struggle to incorporate directed interactions. This perspective presents a comprehensive roadmap for understanding balance in signed directed networks, extending traditional balance theory to account for directed interactions. Balance is indicated by the enrichment of higher-order patterns like triads compared to an adequate null model, where the network is randomized with some key aspects being preserved. Finding appropriate null models has been a challenging task even without considering directionality, which largely expands the space of potential null models. Recently, it has been shown that in the undirected case both the network topology and the signed degrees serve as key factors to preserve. Therefore, we introduce a maximally constrained null model that preserves the directed topology as well as node-level features given by the signed unidirectional, reciprocated, and conflicting node degrees. Our null model is based on the maximum-entropy principle and reveals consistent patterns across large-scale social networks. We also consider directed generalizations of balance theory and find that the observed patterns are well aligned with two proposed directed notions of strong balance. Our approach not only unveils balance in signed directed networks but can also serve as a starting point towards generative models of signed directed social networks, advancing our understanding of complex social systems and their dynamics.