KSLCore/ksl.simopt.solvers.algorithms.bo/Matern52Kernel

Matern52Kernel

class Matern52Kernel(signalVariance: Double, lengthScales: DoubleArray) : StationaryKernel(source)

The Matern 5/2 kernel with ARD length scales. It is less smooth than the RBF kernel (twice mean-square differentiable) and is often a more realistic choice for response surfaces.

cov = signalVariance * (1 + sqrt(5)*r + (5/3)*r^2) * exp(-sqrt(5)*r), where r is the length-scaled Euclidean distance.

Constructors

Matern52Kernel

constructor(signalVariance: Double, lengthScales: DoubleArray)

constructor(problemDefinition: ProblemDefinition, signalVariance: Double = 1.0)

Functions

cov

open override fun cov(x1: DoubleArray, x2: DoubleArray): Double

The prior covariance between the responses at x1 and x2.

toString

open override fun toString(): String