automultinomial (version 1.0.0)

uncenteredLogLikGrad: Uncentered gradient

Description

Takes coefficient values and a design matrix and computes the gradient of the loglikelihood

Usage

uncenteredLogLikGrad(theta, X, z, A, constraint, thetaInds, innerIndices)

Arguments

theta

matrix of coefficients

X

an nxp design matrix

z

an nxk response matrix

A

list of adjacency matrices/list containing adjacency matrix

constraint

constraint on eta

thetaInds

relevant indices

innerIndices

the (1-dim.) location/index for internal points in the grid

Value

a gradient vector

Examples

# NOT RUN {
#gradient=uncenteredLogLikGrad(theta,X,z,A,constraint,thetaInds)
# }