Community Detection in Signed Networks Based on Normalized Laplacian Matrix and Convex Non-negative Matrix Factorization

Signed networks can intuitively and flexibly describe the positive and negative interactions among entities in a complex network. Community structure detection is a basic research topic of signed network analysis, which is of theoretical significance and practical value to reveal the organizational...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 2nd International Conference on Automation, Robotics and Computer Engineering (ICARCE) s. 1 - 5
Hlavní autoři: Zhang, Yibo, Nan, Tian, Zhang, Zhaozhi
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 14.12.2023
Témata:
Community detection
Complex networks
Detection algorithms
Image edge detection
Laplace equations
Matrix decomposition
Network analyzers
Noise
Non-negative matrix factorization
Normalized Laplacian matrix
Signed networks
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Signed networks can intuitively and flexibly describe the positive and negative interactions among entities in a complex network. Community structure detection is a basic research topic of signed network analysis, which is of theoretical significance and practical value to reveal the organizational structure, functional structure, and individual behavior of complex systems. However, it is still a challenging task to accurately detect the community structure in a signed network. In this paper, we design a community detection algorithm based on normalized Laplacian matrix and convex non-negative matrix factorization for signed networks. The proposed algorithm applies convex non-negative matrix decomposition to a signed network, and decomposes the adjacency matrix of the network into a community membership matrix and a node weight matrix. Meanwhile, the algorithm introduces normalized Laplacian graph regularization terms to reduce noise generated during the decomposition process. Experimental results verify that the proposed algorithm is effective in community detection.
DOI:10.1109/ICARCE59252.2024.10492483