The Artificial Benchmark for Community Detection with Outliers and Overlapping Communities (ABCD+$o^2$)

2025-06-05

AI Summary

Existing benchmark graph models (e.g., LFR) lack analytical tractability, fail to model outliers, and do not support overlapping communities—limiting their fidelity for evaluating community detection in realistic, noisy, and overlapping settings. Method: We propose a novel stochastic graph generation model grounded in the ABC (Affiliation-Block-Community) framework. It jointly models power-law degree distributions, power-law community size distributions, outliers, and overlapping communities, enabling closed-form analysis and precise parameter control. Innovations include an extended membership sampling mechanism and a principled outlier injection strategy. Contribution/Results: This is the first model to simultaneously capture all four key structural features of real-world networks. Generated graphs exhibit high structural fidelity and efficient generation scalability, significantly outperforming LFR in both realism and computational efficiency. The model establishes a more robust, fair, and reproducible benchmark for evaluating community detection algorithms under noisy and overlapping conditions.

Technology Category

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

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

The Artificial Benchmark for Community Detection (ABCD) graph is a random graph model with community structure and power-law distribution for both degrees and community sizes. The model generates graphs similar to the well-known LFR model but it is faster, more interpretable, and can be investigated analytically. In this paper, we use the underlying ingredients of the ABCD model, and its generalization to include outliers (ABCD+$o$), and introduce another variant that allows for overlapping communities, ABCD+$o^2$.

Problem

• Extend ABCD model to include overlapping communities
• Improve community detection with outlier handling
• Provide faster and more interpretable graph generation

Innovation

• Extends ABCD model with overlapping communities
• Enhances interpretability and analytical investigation
• Faster generation than LFR benchmark model