wschisq_utils function

Utilities for weighted sums of non-central chi squared random variables

Simulation from a weighted sum of non-central chi squared random variables and Monte Carlo approximation of its distribution function.

r_wschisq_Cpp(n, weights, dfs, ncps)

p_wschisq_MC(x, weights = 0L, dfs = 0L, ncps = 0L, M = 10000L,
  sample = 0L, use_sample = FALSE, x_sorted = FALSE)

Arguments

• n: sample size.

• weights: vector with the positive weights of the sum. Must have the same length as dfs.

• dfs: vector with the positive degrees of freedom of the chi squared random variables. Must have the same length as weights.

• ncps: non-negative non-centrality parameters. A vector with the same length as weights.

• x: vector of quantiles.

• M: number of Monte Carlo samples for approximating the distribution. Defaults to 1e4.

• sample: if use_sample = TRUE, the Monte Carlo sample to approximate the distribution. If not, it is computed internally. Defaults to 1e4.

• use_sample: use the already computed sample? If FALSE

  (default), sample is computed internally.

Returns

• r_wschisq_Cpp: a matrix of size c(n, 1)

  containing a random sample.

• p_wschisq_MC: a matrix of size c(nx, 1)

  with the evaluation of the distribution function at x.