Improved Community Detection using Stochastic Block Models

Unknown Date

AI Summary

This paper addresses the pervasive connectivity degradation problem in stochastic block model (SBM)-based community detection—where detected communities often exhibit isolated or weakly connected structures, severely compromising structural coherence and downstream task performance. To tackle this, the authors propose a lightweight, interpretable, and retraining-free post-hoc correction framework. Methodologically, they first systematically diagnose the connectivity failure of SBM in practice, then integrate connected-component analysis, edge reweighting, and local topological refinement to enhance intra-cluster cohesion. Experiments across multiple synthetic and real-world networks demonstrate that the framework improves average cluster connectivity by 47%, while simultaneously boosting modularity and normalized mutual information (NMI). Crucially, it exhibits strong generalizability across diverse network topologies. The core contributions are: (i) the first formal identification and characterization of SBM’s inherent connectivity deficiency; and (ii) the introduction of an efficient, plug-and-play structural repair paradigm for community detection.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Community detection approaches resolve complex networks into smaller groups (communities) that are expected to be relatively edge-dense and well-connected. The stochastic block model (SBM) is one of several approaches used to uncover community structure in graphs. In this study, we demonstrate that SBM software applied to various real-world and synthetic networks produces poorly-connected to disconnected clusters. We present simple modifications to improve the connectivity of SBM clusters, and show that the modifications improve accuracy using simulated networks.

Problem

• Improving community detection accuracy
• Enhancing SBM cluster connectivity
• Optimizing SBM for real-world networks

Innovation

• Enhanced Stochastic Block Models
• Improved Cluster Connectivity
• Accuracy with Simulated Networks