GLMGraph: Fit a graphical model nodewise via generalized linear...

In hwang655/HUG: Dependent Graphical Model Estimation

Description

Fit a graphical model nodewise via generalized linear regression

Usage

GLMGraph(Y.r, Y.p, Z = NULL, nlambda = 50, family = "gaussian", intercept = T, lambda.min.ratio = 0.01, weight.zero = 1, hetero.penalty = F)

Arguments

Y.r	response of nodewise regressions, an nxp matrix
Y.p	predictors of nodewise regressions, a matrix of the same dimension as Y.r
Z	extra predictor matrix for all nodewise regressions
nlambda	number of lambda values on grid (default 50)
family	"gaussian" or "poisson"
intercept	whether to include an intercept term in each nodewise regressions
lambda.min.ratio	ratio between the minimum and the maximum of the lambda sequence
weight.zero	Weight of zero-response samples in nodewise regressions

Value

lambda	the lambda sequence used in nodewise regressions
graphs	a sequence of estimated graphs
coef.opt	a pxp matrix of EBIC-selected coefficient estimates of all nodewise regressions
coef.aic	a pxp matrix of AIC-selected coefficient estimates of all nodewise regressions

Author(s)

Jianyu Liu

See Also

hugeGraph, MPoisGraph, MHurdGraph

hwang655/HUG documentation built on Jan. 31, 2021, 2:54 a.m.