pmvnorm: Distribution function of the multivariate normal distribution for arbitrary limits

Description

This function computes the distribution function of a multivariate normal distribution vector for an arbitrary rectangular region [lb, ub]. pmvnorm computes an estimate and the value is returned along with a relative error and a deterministic upper bound of the distribution function of the multivariate normal distribution. Infinite values for vectors \(u\) and \(l\) are accepted. The Monte Carlo method uses sample size \(n\): the larger the sample size, the smaller the relative error of the estimator.

Usage

pmvnorm(
  mu,
  sigma,
  lb = -Inf,
  ub = Inf,
  B = 10000,
  type = c("mc", "qmc"),
  check = TRUE,
  ...
)

Arguments

mu: vector of location parameters
sigma: covariance matrix
lb: vector of lower truncation limits
ub: vector of upper truncation limits
B: number of replications for the (quasi)-Monte Carlo scheme
type: string, either of mc or qmc for Monte Carlo and quasi Monte Carlo, respectively
check: logical; if TRUE (default), the code checks that the scale matrix sigma is positive definite and symmetric
...: additional arguments, currently ignored.

Author

Zdravko I. Botev, Leo Belzile (wrappers)

References

Z. I. Botev (2017), The normal law under linear restrictions: simulation and estimation via minimax tilting, Journal of the Royal Statistical Society, Series B, 79 (1), pp. 1--24.

Examples

Run this code

#From mvtnorm
mean <- rep(0, 5)
lower <- rep(-1, 5)
upper <- rep(3, 5)
corr <- matrix(0.5, 5, 5) + diag(0.5, 5)
prob <- pmvnorm(lb = lower, ub = upper, mu = mean, sigma = corr)
stopifnot(pmvnorm(lb = -Inf, ub = 3, mu = 0, sigma = 1) == pnorm(3))