Improving the Accuracy of Community Detection on Signed Networks via Community Refinement and Contrastive Learning
This work addresses the challenge of reliable community detection for researchers and practitioners analyzing signed networks, but it is incremental as it builds on existing methods with a refinement approach.
The paper tackles the problem of inconsistent community detection on signed networks due to noisy edge signs by proposing ReCon, a model-agnostic post-processing framework that refines communities through iterative steps including contrastive learning, resulting in enhanced accuracy across synthetic and real-world networks.
Community detection (CD) on signed networks is crucial for understanding how positive and negative relations jointly shape network structure. However, existing CD methods often yield inconsistent communities due to noisy or conflicting edge signs. In this paper, we propose ReCon, a model-agnostic post-processing framework that progressively refines community structures through four iterative steps: (1) structural refinement, (2) boundary refinement, (3) contrastive learning, and (4) clustering. Extensive experiments on eighteen synthetic and four real-world networks using four CD methods demonstrate that ReCon consistently enhances community detection accuracy, serving as an effective and easily integrable solution for reliable CD across diverse network properties.