createLocScaleGrid: Create a centered and scaled sparse integration grid

In bisque: Approximate Bayesian Inference via Sparse Grid Quadrature Evaluation (BISQuE) for Hierarchical Models

Description

Enhances mvQuad::createNIGrid by shifting and scaling a sparse integration grid, and evaluating the weight function at each of the grid nodes.

Usage

createLocScaleGrid(
  mu = 0,
  prec = 1,
  level = 2,
  quadError = FALSE,
  prec.chol = chol(prec)
)

Arguments

mu	location at which grid should be centered
prec	"precision matrix" associated with the integration grid. When building a sparse integration grid for a density, prec is often the negative of the hessian at the mode.
level	accuracy level. This is typically number of grid points for the underlying 1D quadrature rule. [description from mvQuad::createNIGrid]
quadError	provide additional information about the grid points and integration weights for the quadrature rule with level-1. This information can facilitate approximating the quadrature error.
prec.chol	Upper-triangular Cholesky decomposition of precision matrix.

See Also

mvQuad::createNIGrid

Examples

g = createLocScaleGrid(mu = c(1,0), prec = diag(c(1,.5)), level = 2 )

bisque documentation built on March 26, 2020, 7:27 p.m.