VeccTMVN (version 1.3.1)

Vecc_reorder: Univariate ordering under Vecchia approximation

Description

Univariate ordering under Vecchia approximation

Usage

Vecc_reorder(
  a,
  b,
  m,
  locs = NULL,
  covName = NULL,
  covParms = NULL,
  covMat = NULL
)

Value

new order, nearest neighbor matrix, and coefficient matrix

Arguments

a

lower bound vector for TMVN

b

upper bound vector for TMVN

m

Vecchia conditioning set size

locs

location (feature) matrix n X d

covName

covariance function name from the `GpGp` package

covParms

parameters for `covName`

covMat

dense covariance matrix, not needed when `locs` is not null

Examples

library(lhs)
library(GpGp)
library(VeccTMVN)
set.seed(123)
n <- 100
m <- 5
locs <- lhs::geneticLHS(n, 2)
covparms <- c(1, 0.1, 0)
cov_name <- "matern15_isotropic"
cov_mat <- get(cov_name)(covparms, locs)
a <- rep(-Inf, n)
b <- runif(n)
odr_TN <- TruncatedNormal::cholperm(cov_mat, a, b)$perm
rslt <- Vecc_reorder(a, b, m,
  locs = locs, covName = cov_name,
  covParms = covparms
)
# compare order
cat(rslt$order)
cat(odr_TN)