Description Usage Arguments Value Author(s) References Examples

Fits a balanced stochastic block model to an adjacency matrix using SDP-1. The function implements a first-order ADMM solver for SDP-1.

1 | ```
sdp1_admm(As, K, opts)
``` |

`As` |
a binary adjacency matrix. |

`K` |
number of communities (or blocks). |

`opts` |
a list containing options. Pass the empty list, that is, "list()", to use the default values. (See examples.) |

A list containing the following items:

`X` |
the estimated cluster matrix. |

`delta` |
a vector of norm differences between consecutive cluster matrices at each step of the ADMM iteration. |

`T_term` |
number of actual iterations performed. |

Arash A. Amini

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

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | ```
# Create a simple blkmodel with K=3 communities each of size m=20
blkmodel <- list(m=20, K=3, p=.9, q=.4)
blkmodel <- within(blkmodel, {
n <- m*K
M <- kronecker(matrix(c(p,q,q,q,p,q,q,q,p),nrow=3),matrix(1,m,m))
As <- 1*(matrix(runif(n^2),nrow=n) < M)
})
# Call sdp1_admm with options:
# rho the ADMM parameter,
# T maximum number of iteration
# tol tolerance for norm(X_{t+1} - X_t)
# report_interval how many iteration between reporting progress
sdp.fit <- with(blkmodel,
sdp1_admm(as.matrix(As), K, list(rho=.1, T=10000, tol=1e-5, report_interval=100)))
# plot the adjacency matrix and the estimated cluster matrix
par(mfrow=c(1,2))
image(blkmodel$As)
image(sdp.fit$X)
``` |

sbmSDP documentation built on May 2, 2019, 9:35 a.m.

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.