The paper introduces the Signed Generalized Random Dot Product Graph (SGRDPG) model, extending it to a multiplex version. It emphasizes the importance of maintaining edge signs for better precision in estimation and clustering. The study explores various stochastic network models, focusing on relationships between nodes across layers. By employing novel algorithms, the paper ensures consistent clustering of layers and high subspace estimation accuracy. The research demonstrates theoretical guarantees through numerical simulations and real data examples, showcasing the benefits of signed networks in analysis. The content delves into balancing theory in signed networks and its implications on clusterability. Various authors' perspectives on balance theory are discussed, highlighting contrasting conclusions drawn from real-world network studies.
Sang ngôn ngữ khác
từ nội dung nguồn
arxiv.org
Thông tin chi tiết chính được chắt lọc từ
by Marianna Pen... lúc arxiv.org 03-18-2024
https://arxiv.org/pdf/2402.10242.pdfYêu cầu sâu hơn