Community Detection in Signed Networks

Community detection in signed networks seeks to partition nodes into cohesive groups by accounting for both positive links, which signify affinity or cooperation, and negative links, which indicate antagonism or conflict. Unlike unsigned networks, where communities are defined solely by dense clusters of positive interactions, signed networks require methods that balance attraction and repulsion. This duality poses unique challenges: negative ties may fragment communities or highlight boundaries between antagonistic factions, while positive ties promote cohesion within groups. Advances in this field have broad implications for understanding social alliances and rivalries, metabolic and neural interactions, and collaborative structures in organisations. By integrating negative links, researchers can reveal hidden structural patterns, refine the detection of overlapping or nested communities, and improve the stability and interpretability of partitions across multiple scales.

Research from Nature Portfolio

Recent studies have extended fundamental frameworks to the signed setting, shedding light on the role of negative ties in shaping mesoscopic structure. One line of work adapts the map equation—originally conceived for flow-based community detection—by introducing penalties for negative links that weaken intra-community flow and increase the likelihood of escape events. This approach also generalises the Constant Potts Model to signed networks, producing a multiscale spectrum that quantifies how informative negative ties are at different resolutions and identifies their topological placement among positive clusters. A complementary development employs random-walk dynamics to establish community memberships: initial seeds are chosen based on local degree maxima, then transition probabilities driven by positive links are balanced against repulsion probabilities from negative links. Nodes join existing communities when affinity outweighs enmity, otherwise they seed new communities, and a subsequent optimisation step merges similar groups. Both lines of inquiry demonstrate that explicitly modelling negative relations enhances the detection of meaningful partitions in synthetic benchmarks and empirical datasets.

Community Detection in Signed Networks publication trend

The graph below shows the total number of articles in community detection in signed networks across all publications each year (not limited to Nature Index journals).

Technical terms

Signed network: A graph in which edges carry positive or negative values to indicate cooperative or antagonistic relationships.

Community detection: The process of identifying groups of nodes that are more densely connected internally than with the rest of the network.

Modularity: A quality function that measures the strength of division of a network into communities by comparing observed link density to a random baseline.

Map equation: An information-theoretic objective that models random-walk flows to reveal community structure by minimising a description length.

Potts model: A statistical-physics framework for partitioning networks, generalised to control resolution via a tunable parameter balancing edge density.

Social balance theory: A principle positing that triadic relationships in signed networks tend toward balanced configurations, with an even number of negative ties promoting stability.

Overlapping communities: Groups in which nodes may belong to more than one community simultaneously, reflecting multifaceted affiliations.

References

1. Community Detection in Signed Networks: the Role of Negative ties in Different Scales. Scientific Reports (2015).
2. Overlapping community detection in networks with positive and negative links. Journal of Statistical Mechanics Theory and Experiment (2014).
3. An algorithm based on positive and negative links for community detection in signed networks. Scientific Reports (2017).
4. A Memetic Algorithm for Community Detection in Signed Networks. IEEE Access (2020).
5. Modularity, balance, and frustration in student social networks: The role of negative relationships in communities. PLOS ONE (2022).