mvdn.cov: MVDN: Matrix-variate Differential Network Analysis.

In jijiadong/MVDN: Matrix-variate Differential Network Analysis

Description

Matrix-Variate Differential Network (MVDN) model which is particularly useful in modelling connectivity alteration for matrix-variate data such as in brain connectivity analysis.

Usage

mvdn.cov(S1, S2, n1, n2, lambda = NULL, nlambda = 10, method = c("none", "aic", "bic", "nbic"), lambda.min.ratio = NULL, rho = NULL, shrink = NULL, prec = 0.001)

Arguments

S1	A p*p matrix.
S2	A p*p matrix.
n1	Sample size of group1.
n2	Sample size of group2.
lambda	The tuning parameter of lasso penalty. user-supplied lambda sequence; default is NULL.
nlambda	The number of lambda values, default is 20.
method	The method used in the lambda selection.
lambda.min.ratio	The ratio used to get the min lambda if lambda is NULL.
rho	The parameter in augmented Lagrange method.
shrink	The parameter shrink the lambda.
prec	The parameter shrink the lambda

Value

mvdn	differential network.
lambda	the actual sequence of lambda values used.
nlambda	the number of lambda values used.
opt	indicating which one is the optimized lambda based on different norms ("max": max norm, "1": element-wise, "L1": matrix L1 max norm, "Spectral", "Frobenius", "Nuclear").

See Also

mvdn

jijiadong/MVDN documentation built on Sept. 6, 2020, 7:15 p.m.