Improving community detection via community association strength scores

2025-01-29

AI Summary

Traditional community detection algorithms impose hard partitioning, forcing each node into a single community while neglecting overlapping memberships and outliers. To address this limitation, we propose a lightweight Community Association Strength (CAS) post-processing framework that quantifies node–community affiliations via three orthogonal scores: neighborhood overlap degree, intra-community cohesion, and inter-community sparsity. CAS operates without re-partitioning the network and enables unsupervised identification of multi-membership nodes and outliers, thereby relaxing the hard-clustering constraint. Evaluated on multiple benchmark datasets, CAS consistently improves modularity and normalized mutual information (NMI). It accurately detects 5–12% outlier nodes and 18–35% overlapping members. To our knowledge, CAS is the first method to systematically unify overlapping community detection and outlier identification within a single, efficient post-processing paradigm.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Reasoning under Uncertainty: Causality

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: Community question answering
• Social Networks and Social Media: Computational social science

Abstract

Community detection methods play a central role in understanding complex networks by revealing highly connected subsets of entities. However, most community detection algorithms generate partitions of the nodes, thus (i) forcing every node to be part of a community and (ii) ignoring the possibility that some nodes may be part of multiple communities. In our work, we investigate three simple community association strength (CAS) scores and their usefulness as post-processing tools given some partition of the nodes. We show that these measures can be used to improve node partitions, detect outlier nodes (not part of any community), and help find nodes with multiple community memberships.

Problem

• Community Detection
• Multi-membership
• Node Classification

Innovation

• Community Detection
• Multi-membership Nodes
• Network Analysis Optimization