dtmvnormMvImportSampling: Truncated Multivariate Gaussian density

In pasudyan/ConstrMixMod: MCMC methods for flexible mixture modeling on constrained spaces

Description

This function evaluates the truncated multivariate Gaussian density using an important sampling procedure as the normalizing constant

Usage

dtmvnormMvImportSampling(X, mn, sig, constr, numSamp = 10000, log = T)

Arguments

X	a vector of observations where each row is the dimension and each column is the independent observations
mn	mean of the multivariate Gaussian density
sig	covariance matrix of the multivariate Gaussian density
constr	a function that describes the domain or constraint of the samples
numSamp	number of samples in the important sampling procedure
log	whether log value is returned (default TRUE)

Value

value of density

Examples

genconst <- function(u1, u2, l1, l2) { in_set <- function(X) {
 if (nrow(X) == 1) X = t(X)
 X[1,] > l1 & X[1,] < u1 & X[2,] > l2 & X[2,] < u2 & X[3,] > l1 &
 X[3,] < u1 & X[4,] > l2 & X[4,] < u2 }}
constr  <- genconst(u1 = 0, u2 = 0, l1 = 1, l2 = 1)
dtmvnormMvImportSampling(X = mvnfast::rmvn(5, mu = rep(0,4), sigma = diag(0.1, 4)), mn = rep(0, 4), sig = diag(0.1,4), constr, numSamp = 1000, log = F)

pasudyan/ConstrMixMod documentation built on May 10, 2019, 8:26 a.m.