rstnorm_ineq_cork: Sample from multivariate standard normal distribution with...
In dcbdan/s525: Implementation of common Bayesian models

Description Usage Arguments Details Examples

Sample from multivariate normal distribution with mean 0 and covariance matrix I_d with the constraint that M^t z ≤q m where M is given by ineq.mat and m is given by ineq.vec

rstnorm_ineq_cork = function(
  n,
  ineq.mat,
  ineq.vec,
  thin = 1,
  burn = 0,
  init = NULL)

`n`	number of elements to generate
`ineq.mat`	Inequality matrix
`ineq.vec`	Inequality vector
`thin`	Take every thinth value from the gibbs sampler
`burn`	Discard the first burn elements generated
`init`	initial z value. If not given, an SVD of ineq.mat happens.

See Schmidt, Mikkel. "Linearly constrained bayesian matrix factorization for blind source separation." Advances in neural information processing systems. 2009.

# sample on a cube!
M = rbind(c(-1,1,0,0),c(0,0,-1,1))
xmin = 1
xmax = 100
ymin = 1
ymax = 100
m = c(-xmin, xmax, -ymin, ymax)
n = 1000
rr = rstnorm_ineq_cork(n, M, m)
plot(rr[,1], rr[,2])

# sample on a triangle and have the third dimension truncated positive
M = cbind(c(-1, 0, 0), c(0, -1, 0), c(1, 1, 0), c(0, 0, -1))
m = c(0, 0, 3, 0)
n = 10000
rr = rstnorm_ineq_cork(n, M, m)
plot(rr[,1], rr[,2])
hist(rr[,3])