rnorm_qinv_l: Sample from multivariate normal distribution

In dcbdan/s525: Implementation of common Bayesian models

Description

Sample from multivariate normal distribution with mean Q^{-1}l and covariance matrix Q^{-1}.

Usage

rnorm_qinv_l <- function(
  n,
  Q,
  l,
  L)

Arguments

n	number of elements to generate
Q	p by p precision matrix.
l	p by 1 vector
L	NULL by default. If not null, Q is ignored and assumed to be LL^t.

Details

The algorithm is as follows

1. Cholesky decomposition of Q into LL^t. (This step is skipped if L is passed in).

2. Sample z from rnorm(p). Let y = Lz + l.

3. Solve for x in LL^tx = y and return.

Value

x	An p x 1 vector if n=1 otherwise a n by p matrix

dcbdan/s525 documentation built on May 19, 2019, 10:48 p.m.