Gaussian Processes

Tools for defining and working with Gaussian processes.

This group contains classes for defining Gaussian processes and performing Gaussian process regression.

const unsigned numFields = 1;   // The number of components in the GP predictions
const unsigned dim       = 1;   // The number of predictor variables (e.g., location)
const unsigned numobs    = 100; // The number of observations
const unsigned numPred   = 10;  // The number of locations where we want to evaluate the GP
 
Eigen::MatrixXd trainLocs(dim,       numObs);
Eigen::MatrixXd trainData(numFields, numObs);
 
Eigen::MatrixXd predLocs(dim, numPred);
 
// ... Fill in the training locations and data
// ... Fill in the prediction locations
 
// Create the covariance kernel
const double lengthScale = 0.35;
const double variance    = 2.0;
 
auto kernel = SquaredExpKernel(dim, variance, lengthScael);
 
// Create the Mean function
ConstantMean mean(dim, 1);
 
// Create the Gaussian Process
auto gp = GaussianProcess(mean, kernel);
 
// Make a prediction at new locaitons
Eigen::MatrixXd postMean, postCov;
std::tie(postMean, postCov) = gp.Predict(predLocs);

Modules
	Covariance Kernels
	Covariance kernels for defining Gaussian Processes.
	Mean Functions
	GP_Observations
	Tools for defining linear observations of Gaussian Processes.

Classes
class	muq::Approximation::GaussianProcess
class	muq::Approximation::KarhunenLoeveExpansion
	Used to compute and evaluate the Karhunen-Loeve decomposition of a zero mean Gaussian process. @sealso SeperableKarhunenLoeve. More...
class	muq::Approximation::SeparableKarhunenLoeve
	Implements KL expansions that take advantage of separable structure in both the domain and covariance kernel. More...

Functions
template<typename MeanType , typename KernelType >
GaussianProcess	muq::Approximation::ConstructGP (MeanType const &mean, KernelType const &kernel)

Function Documentation

ConstructGP()

template<typename MeanType , typename KernelType >

GaussianProcess muq::Approximation::ConstructGP	(	MeanType const &	mean,
		KernelType const &	kernel
	)

Definition at line 389 of file GaussianProcess.h.