quick_C: Covariance of kernel computations

In activegp: Gaussian Process Based Design and Analysis for the Active Subspace Method

Covariance of kernel computations

Description

Computes Int(kappa_i(X, design) . kappa_j(design, X)). This function is preferred for initialization

Usage

quick_C(measure, design, Ki, Kir, theta, xm, xv, ct, verbose)

Arguments

measure	An integer giving the measure of integration. 0 for Lebesgue/Uniform on [0,1]^m, 1 for (truncated) Gaussian on [0,1]^m, 2 for Gaussian on R^m.
design	matrix of design points
Ki	The inverse covariance matrix
Kir	The inverse covariance matrix times the response.
theta	lengthscales
xm	The mean vector associated with the Gaussian measure. Ignored if uniform.
xv	The variance vector associated with the Gaussian measure (diagonal of covariance matrix, vars assumed independent). Ignored if uniform.
ct	Covariance type, 1 means Gaussian, 2 means Matern 3/2, 3 means Matern 5/2

Value

The matrix representing the result of the integration.

activegp documentation built on June 28, 2022, 1:05 a.m.