TruncatedNormal

Truncated multivariate Normal and Student distributions

A collection of functions to deal with the truncated univariate and multivariate normal and Student distributions, described in Botev (2017) and Botev and L'Ecuyer (2015).

Main features are

• simulation from multivariate truncated Normal and student distributions using an accept-reject algorithm based on minimax exponential tilting.
• (quasi) Monte-Carlo estimation of the distribution function using separation-of-variables together with exponential tilting for provable performances and theoretical upper bound on the error.
• Cholesky decomposition using the reordering algorithm of Gibson, Glasbey and Elston (1994).

To install the latest version from CRAN, use

install.packages("TruncatedNormal")

or else install the latest development from Github via

# devtools::install_github("lbelzile/TruncatedNormal")
install.packages("TruncatedNormal", repos = c('https://lbelzile.r-universe.dev', 'https://cloud.r-project.org'))