gaussianprocess

Documentation

Index

• type Regressor
  • func NewRegressor(kernel kernels.Kernel) *Regressor
  • func (m *Regressor) Fit(X, Y mat.Matrix) base.Fiter
  • func (m *Regressor) GetNOutputs() int
  • func (m *Regressor) IsClassifier() bool
  • func (m *Regressor) LogMarginalLikelihood(Theta mat.Matrix, evalGradient bool) (lml float64, grad []float64)
  • func (m *Regressor) Predict(X mat.Matrix, Y mat.Mutable) *mat.Dense
  • func (m *Regressor) PredictEx(X mat.Matrix, Y mat.Mutable, returnStd, returnCov bool) (*mat.Dense, *mat.DiagDense, *mat.Dense)
  • func (m *Regressor) PredicterClone() base.Predicter
  • func (m *Regressor) Score(X, Y mat.Matrix) float64

Types

type Regressor

type Regressor struct {
	kernels.Kernel
	Alpha []float64
	// Optimizer is always optimize.LBFGS
	NRestartsOptimizer int
	NormalizeY         bool
	// copy_X_train is always true
	base.RandomState
	Xtrain                     *mat.Dense
	Ytrain                     *mat.Dense
	YtrainMean                 *mat.Dense
	KernelOpt                  kernels.Kernel
	L                          *mat.Cholesky
	LogMarginalLikelihoodValue float64
}

Regressor ...

func NewRegressor

func NewRegressor(kernel kernels.Kernel) *Regressor

NewRegressor ...

func (*Regressor) Fit

func (m *Regressor) Fit(X, Y mat.Matrix) base.Fiter

Fit Gaussian process regression model

func (*Regressor) GetNOutputs

func (m *Regressor) GetNOutputs() int

GetNOutputs returns Y columns count

func (*Regressor) IsClassifier

func (m *Regressor) IsClassifier() bool

IsClassifier returns false

func (*Regressor) LogMarginalLikelihood

func (m *Regressor) LogMarginalLikelihood(Theta mat.Matrix, evalGradient bool) (
	lml float64, grad []float64,
)

LogMarginalLikelihood returns log-marginal likelihood of theta for training data

func (*Regressor) Predict

func (m *Regressor) Predict(X mat.Matrix, Y mat.Mutable) *mat.Dense

Predict using the Gaussian process regression model

func (*Regressor) PredictEx

func (m *Regressor) PredictEx(X mat.Matrix, Y mat.Mutable, returnStd, returnCov bool) (*mat.Dense, *mat.DiagDense, *mat.Dense)

PredictEx predicts using the Gaussian process regression model, returning Ymean and std or cov

func (*Regressor) PredicterClone

func (m *Regressor) PredicterClone() base.Predicter

PredicterClone clones Predicter (for KFold...)

func (*Regressor) Score

func (m *Regressor) Score(X, Y mat.Matrix) float64

Score returns R2 score