add.kernel: Add a Kernel Class to a Model Tree

In mattdneal/gaussianProcess: Gaussian Process Fitting and Prediction

Description

Add a Kernel Class to a Model Tree

Usage

add.kernel(orig.model.tree, kernel.class.name, kernel)

Arguments

orig.model.tree	a ModelTree object
kernel.class.name	the name of the kernel class (arbitrary, must not match the class name of any kernel already in the model tree).
kernel	either an R function defining a kernel, or a string indicating a built-in kernel type
hyper.param.names	a vector containing the names of the kernel hyperparameters
kernel.additional.params	a list containing any additional kernel parameters
kernel.grad	either an R function returning the gradient of kernel, or NULL. If kernel is a string specifying a built-in kernel, kernel.grad should be NULL. If kernel is an R function and kernel.grad is NULL, grad will be used to derive a numerical gradient.

Value

a ModelTree object

Examples

mt <- create.model.tree()
mt <- add.kernel(mt, "SE", "squaredExponential", c("l"))

mattdneal/gaussianProcess documentation built on May 21, 2019, 12:58 p.m.