I decided to break off from matrixdist the portion dedicated to the Wishart-related functions.
They are self-contained and don’t really exist on their own elsewhere (there are a few that include
them along with a lot of other functionality, some in C++ but a lot in R), so it’s good to have a little
package that offers them on their own. Not everybody would want or need matrixdist and it’s good to
offer the option without polluting the NAMESPACE.
The package offers:
- Sampling from the Cholesky factorization of a Wishart distribution.
- Sampling from an inverse Wishart distribution.
- Sampling from the Cholesky factorization of an inverse Wishart distribution.
- Density function for the Wishart and inverse Wishart distributions.
- Multivariate gamma and digamma functions.
The sampling functions are in C and are based on the function stats::rWishart() so they are
roughly as fast.
set.seed(20180226)
sigma <- rWishart(1, 10, diag(6))[,,1]
args <- list(n = 10000, df = 10, Sigma = sigma)
library('CholWishart')
library('microbenchmark')
library('ggplot2')
results <- microbenchmark(
A <- do.call("rWishart", args),
A <- do.call("rCholWishart",args),
A <- do.call("rInvWishart",args),
A <- do.call("rInvCholWishart",args)
)
autoplot(results)+theme_bw()## Coordinate system already present. Adding new coordinate system, which will replace the existing one.
Check out the package on CRAN or install:
install.packages('CholWishart')or get the lastest development version:
devtools::install_github("gzt/CholWishart")