lmn_prof: Profile loglikelihood for the LMN model.

In mlysy/LMN: Inference for Linear Models with Nuisance Parameters

Profile loglikelihood for the LMN model.

Description

Calculate the loglikelihood of the LMN model defined in lmn_suff() at the MLE Beta = Bhat and Sigma = Sigma.hat.

Usage

lmn_prof(suff, noSigma = FALSE)

Arguments

suff	An object of class lmn_suff (see lmn_suff()).
noSigma	Logical. If TRUE assumes that Sigma = diag(ncol(Y)) is known and therefore not estimated.

Value

Scalar; the calculated value of the profile loglikelihood.

Examples

# generate data
n <- 50
q <- 2
Y <- matrix(rnorm(n*q),n,q) # response matrix
X <- matrix(1,n,1) # covariate matrix
V <- exp(-(1:n)/n) # diagonal variance specification
suff <- lmn_suff(Y, X = X, V = V, Vtype = "diag") # sufficient statistics

# profile loglikelihood
lmn_prof(suff)

# check that it's the same as loglikelihood at MLE
lmn_loglik(Beta = suff$Bhat, Sigma = suff$S/suff$n, suff = suff)

mlysy/LMN documentation built on March 25, 2022, 11:12 a.m.