randnet (version 0.7)

RightSC: clusters nodes in a directed network by regularized spectral clustering on right singular vectors

Description

community detection by regularized spectral clustering on right singular vectors

Usage

RightSC(A, K, normal = FALSE)

Value

a list of

cluster

cluster labels

loss

the loss of Kmeans algorithm

Arguments

A

adjacency matrix of a directed adjacecy matrix

K

number of communities

normal

indicator. If TRUE, normalization of singular vector rows will be applied, similar to the spectral spherical clustering.

Author

Tianxi Li, Elizaveta Levina, Ji Zhu

Maintainer: Tianxi Li <[e-mail]>

Details

This is essentially the spectral clustering applied on right singular vectors. It can be used to handle directed networks where Aij = 1 if and only if i -> j, and the edges tend to have a missing issue specifically depending on the sender i. More details can be found in Li et. al. (2020).

References

T. Li, E. Levina, and J. Zhu. Community models for networks observed through edge nominations. arXiv preprint arXiv:2008.03652 (2020).

See Also

reg.SP

Examples

dt <- BlockModel.Gen(30,300,K=3,beta=0.2,rho=0)


A <- dt$A


sc <- RightSC(A,K=2)