gp_lml_deriv_helper: Compute the derivative of the log marginal likelihood (helper...

In ebenmichael/gaussianProcess: Fits Gaussian Processes for Regression

Description

Compute the derivative of the log marginal likelihood (helper function)

Usage

gp_lml_deriv_helper(hyper.params, diff.mat, y, K, K.inv, alpha, kernel.deriv,
  noise.var)

Arguments

hyper.params	Hyper parameters as a vector (for rbf c(amplitude, scales))
diff.mat	The result of pdiff(X,X) for input data X
y	The target vector
K	The kernel matrix of X
K.inv	The inverse of K (precomputed)
alpha	An n dimensional vector, alpha = K^-1 y
kernel.deriv	The derivative function for the kernel
noise.var	The variance of the noise around the function

Value

The gradient of the log marginal likelihood at hyper.params

ebenmichael/gaussianProcess documentation built on May 15, 2019, 7:30 p.m.