sbmSDP-package {sbmSDP}

Semidefinite Programming for Fitting Block Models of Equal Block Sizes

Description

An ADMM implementation of SDP-1, a semidefinite programming relaxation of the maximum likelihood estimator for fitting a block model. SDP-1 has a tendency to produce equal-sized blocks and is ideal for producing a form of network histogram approximating a nonparametric graphon model. Alternatively, it can be used for community detection. (This is experimental code, proceed with caution.)

Details

Package:	sbmSDP
Type:	Package
Version:	0.2
Date:	2015-06-18
License:	GPL-3

An ADMM implementation of SDP-1 algorithm for fitting stochastic block models (SBMs). The main function is sdp1_admm.

Author(s)

Arash A. Amini

Maintainer: Arash A. Amini <[e-mail]>

References

On Semidefinite relaxations of the block model by A.A. Amini and E. Levina.