hglasso-package {hglasso}

Fit the hub graphical lasso, hub covariance graph, and hub binary network

Description

This package is called hglasso, for "hub graphical lasso". It implements three methods:hub graphical lasso, hub covariance graph, and hub binary network. All are described in the paper "Learning graphical models with hubs", by Tan et al. (2014).

The main functions are as follows: (1) hglasso (2) hcov (3) hbn

The first function, hglasso, performs hub graphical lasso. The second function, hcov, performs hub covariance graph estimation. The third function, hbn, performs hub binary network estimation.

Details

Package:	hglasso
Type:	Package
Version:	1.2
Date:	2014-08-09
License:	GPL (>=2.0)
LazyLoad:	yes

The package includes the following functinos:

hglasso:	Performs hub graphical lasso
hcov:	Performs hub covariance graph estimation
hbn:	Performs hub binary network estimation
HubNetwork:	Generates inverse covariance matrix or covariance matrix with hubs
binaryMCMC:	Generates samples for binary Ising model via Gibbs sampling
image.hglasso:	Creates image plot of the matrix V and Z
plot.hglasso:	Creates a graphical representation of the estimated matrix Theta
summary.hglasso:	Provides summary for the matrix Theta, Z, and V
hglassoBIC:	Calculate BIC-type criterion for hglasso

Author(s)

Kean Ming Tan and Karthik Mohan

Karthik Mohan implemented the Barzilai-Borwein method for hbn

Maintainer: Kean Ming Tan <[e-mail]>

References

Tan, KM., London, P., Mohan, K., Lee, S-I., Fazel, M., and Witten, D. (2014). Learning graphical models with hubs. Journal of Machine Learning Research 15(Oct):3297-3331.

See Also

hglasso hcov hbn

Examples

##############################################
# Example from Figure 1 in the manuscript
# A toy example to illustrate the results from 
# Hub Graphical Lasso
##############################################
#library(mvtnorm)
#set.seed(1)
#n=100
#p=100

# A network with 4 hubs
#network<-HubNetwork(p,0.99,4,0.1)
#Theta <- network$Theta
#truehub <- network$hubcol
# The four hub nodes have indices 14, 42, 45, 78
#print(truehub)

# Generate data matrix x
#x <- rmvnorm(n,rep(0,p),solve(Theta))
#x <- scale(x)

# Run Hub Graphical Lasso to estimate the inverse covariance matrix
# res1<-hglasso(cov(x),0.3,0.3,1.5)

# print out a summary of the object hglasso
#summary(res1)
# we see that the estimated hub nodes have indices 14, 42, 45, 78
# We successfully recover the 4 hub nodes

# Plot the matrices V and Z 
#image(res1)
#dev.off()
# Plot a graphical representation of the estimated inverse
# covariance matrix --- conditional independence graph
#plot(res1,main="Conditional Independence Graph")