PLordprob: Multivariate Ordered Probit Model via Pairwise Likelihood

Multivariate ordered probit model, i.e. the extension of the scalar ordered probit model where the observed variables have dimension greater than one. Estimation of the parameters is done via maximization of the pairwise likelihood, a special case of the composite likehood obtained as product of bivariate marginal distributions. The package uses the Fortran 77 subroutine SADMVN by Alan Genz, with minor adaptations made by Adelchi Azzalini in his "mvnormt" package for evaluating the two-dimensional Gaussian integrals involved in the pairwise log-likelihood. Optimization of the latter objective function is performed via quasi-Newton box-constrained optimization algorithm, as implemented in nlminb.

AuthorEuloge Clovis Kenne Pagui [aut,cre], Antonio Canale [aut], Alan Genz [ctb], Adelchi Azzalini [ctb]
Date of publication2014-10-07 18:28:10
MaintainerEuloge Clovis Kenne Pagui <kenne@stat.unipd.it>
LicenseGPL-2
Version1.0

View on CRAN

Files in this package

PLordprob
PLordprob/src
PLordprob/src/Makevars
PLordprob/src/cat_pair.c
PLordprob/src/sadmvnt.f
PLordprob/NAMESPACE
PLordprob/R
PLordprob/R/a2delta.R PLordprob/R/delta2a.R PLordprob/R/ordprob.pair.R PLordprob/R/ordprob.mple.R
PLordprob/MD5
PLordprob/DESCRIPTION
PLordprob/man
PLordprob/man/PLordprob-package.Rd PLordprob/man/ordprob.mple.Rd PLordprob/man/delta2a.Rd

Questions? Problems? Suggestions? or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.