New spectral clustering algorithm achieves exact community recovery in bipartite networks

Researchers have developed a spectral clustering algorithm capable of exactly recovering communities in bipartite networks. This algorithm, based on the diagonal-deleted Gram matrix, provides theoretical guarantees for exact recovery under mild conditions, even with unbalanced community sizes or heterogeneous degrees. The method is also effective for degree-corrected stochastic co-blockmodels, maintaining its recovery guarantee with a row-normalized version of the algorithm. Experimental results confirm the theoretical findings. AI

IMPACT Enhances foundational methods for analyzing complex network data, potentially improving recommendation systems and social network analysis.

RANK_REASON Academic paper detailing a new algorithm for community detection in bipartite networks. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv stat.ML →

• arXiv
• degree-corrected stochastic co-blockmodel
• Exact Community Recovery in Bipartite Networks
• Hugging Face
• Social and Information Networks
• spectral clustering
• stochastic co-blockmodel

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →