covariance_kernels (pmrf.covariance_kernels)

Covariance kernels for Gaussian processes.

Useful for discrepancy modeling. See pmrf.discrepancy_models for more details.

Classes

AbstractCovarianceKernel()	Abstract base class for covariance kernel functions.
AutoCrossKernel(auto, cross, num_outputs)	Kernel that routes between a auto-correlation and cross-correlation kernels.
ConstantKernel(variance)	Kernel that returns a constant variance.
Matern32Kernel(lengthscale)	Matérn kernel with nu=3/2.
Matern52Kernel(lengthscale)	Matérn kernel with nu=5/2.
PeriodicKernel(period, lengthscale)	Periodic (Exp-Sine-Squared) kernel.
ProductKernel(k1, k2)	Kernel representing the product of two kernels.
RBFKernel(lengthscale)	Radial Basis Function (Squared Exponential) kernel.
SharedIndependentKernel(base_kernel, ...)	Evaluates a base kernel and broadcasts its output to represent multiple independent dimensions (e.g., real and imaginary parts) withed share hyperparameters.
SumKernel(k1, k2)	Kernel representing the sum of two kernels.
WhiteNoiseKernel(variance)	Kernel representing independent Gaussian noise.
ZeroKernel()	Kernel that always evaluates to zero.