Improving the Accuracy of Community Detection on Signed Networks via Community Refinement and Contrastive Learning

2026-01-23

AI Summary

This work addresses the inconsistency in community detection results caused by noisy or conflicting edge signs in signed networks. To this end, the authors propose ReCon, a model-agnostic post-processing framework that refines existing community assignments through an iterative four-step mechanism involving structural optimization, boundary adjustment, contrastive learning, and clustering. Notably, ReCon is the first to integrate contrastive learning into the post-processing stage of signed network community detection, substantially enhancing both consistency and accuracy. Extensive experiments demonstrate that ReCon consistently improves performance across four state-of-the-art community detection methods on 18 synthetic and 4 real-world signed networks, confirming its effectiveness and broad applicability.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Non-convex Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Social Networks and Social Media: Computational social science
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

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.

Problem

• signed networks
• community detection
• noisy edges
• conflicting edge signs
• community inconsistency

Innovation

• signed networks
• community detection
• contrastive learning
• community refinement
• model-agnostic framework