A header-only R package implementing LinConGauss in C++ and R.
The original Python implementation can be found here. The reference of LinConGauss is this paper. It's very likely this repo will only implement the sampling methods in the near future.
It is not on CRAN and probably won't, to install, use
devtools::install_github("YunyiShen/linconGaussR")An example would be:
my_sample <- linconGauss(100, diag(2),c(0,0),diag(2),c(0,0))
MASS_sample <- MASS::mvrnorm(1000,c(0,0),diag(2))
plot(MASS_sample)
points(my_sample,col = "red")
abline(h=0)
abline(v=0)For developers: This package was written as header-only, with all the sampling method/classes used in the inst/include directory. If you just wish to use the C++ API of this implementation, consider install this package, and add linconGaussR (for sure RcppArmadillo and Rcpp) to your LinkingTo field of the description. The main sampling function is linconGaussR::linconGauss_cpp. Below is a sample implementation calling this method only:
// [[Rcpp::depends(RcppArmadillo)]]
#include <linconGaussR.h>// we call RcppArmadillo
using namespace Rcpp;
using namespace arma;
using namespace std;
using namespace linconGaussR;
//[[Rcpp::export]]
arma::mat linconGauss_cpp(int n, // sample size
arma::mat A, // linear constraint
arma::vec b, // offset of the linear constraint, so that Ax+b>0
arma::mat Sigma, // covariance
arma::vec mu, // mean
arma::vec x_init, // an initial value, necessary
bool intersection=true, // whether to sample from the intersection (otherwise from the joint)
int nskp=5){ // number of sample to skip during iterations
return linconGaussR::linconGauss_cpp(n,A,b,Sigma,mu,x_init,intersection,nskp);
}