Gaussian Mixture Models for Stochastic Block Models with Non-Vanishing\n Noise

2019/11/05 by Heather Mathews, Mathews, Heather, Vaishakhi Mayya +5
Computer Science · Physics and Astronomy · #Bayesian Methods and Mixture Models #Complex Network Analysis Techniques #FOS: Computer and information sciences #Methodology (stat.ME) #Opinion Dynamics and Social Influence

paper · pdf · doi:10.48550/arxiv.1911.01855

openalex publication_date 2019/11/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Community detection tasks have received a lot of attention across statistics,\nmachine learning, and information theory with a large body of work\nconcentrating on theoretical guarantees for the stochastic block model. One\nline of recent work has focused on modeling the spectral embedding of a network\nusing Gaussian mixture models (GMMs) in scaling regimes where the ability to\ndetect community memberships improves with the size of the network. However,\nthese regimes are not very realistic. This paper provides tractable methodology\nmotivated by new theoretical results for networks with non-vanishing noise. We\npresent a procedure for community detection using GMMs that incorporates\ncertain truncation and shrinkage effects that arise in the non-vanishing noise\nregime. We provide empirical validation of this new representation using both\nsimulated and real-world data.\n

Related

• Beyond Asymptotics: Practical Insights into Community Detection in Complex Networks
• Optimal Noise Reduction in Dense Mixed-Membership Stochastic Block Models under Diverging Spiked Eigenvalues Condition
• Exact Recovery of Community Detection in dependent Gaussian Mixture Models
• Community Detection for Contextual-LSBM: Theoretical Limitations of Misclassification Rate and Efficient Algorithms
• Mixture Models and Networks -- Overview of Stochastic Blockmodelling