smile.regression

Class GaussianProcessRegression.Trainer<T>

java.lang.Object

smile.regression.RegressionTrainer<T>

smile.regression.GaussianProcessRegression.Trainer<T>

Enclosing class:

GaussianProcessRegression<T>

public static class GaussianProcessRegression.Trainer<T>
extends RegressionTrainer<T>

Trainer for Gaussian Process for Regression.

Constructor Summary

Constructors

Constructor and Description
GaussianProcessRegression.Trainer(MercerKernel<T> kernel, double lambda) Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
GaussianProcessRegression<T>	train(T[] x, double[] y) Learns a regression model with given training data.
GaussianProcessRegression<T>	train(T[] x, double[] y, T[] t) Learns a Gaussian Process with given subset of regressors.

Methods inherited from class smile.regression.RegressionTrainer

setAttributes

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

GaussianProcessRegression.Trainer

public GaussianProcessRegression.Trainer(MercerKernel<T> kernel,
                                 double lambda)

Constructor.

Parameters:

kernel - the Mercer kernel.

lambda - the shrinkage/regularization parameter.

Method Detail

train

public GaussianProcessRegression<T> train(T[] x,
                                 double[] y)

Description copied from class: RegressionTrainer

Learns a regression model with given training data.

Specified by:

train in class RegressionTrainer<T>

Parameters:

x - the training instances.

y - the training response values.

Returns:

a trained regression model.

train

public GaussianProcessRegression<T> train(T[] x,
                                 double[] y,
                                 T[] t)

Learns a Gaussian Process with given subset of regressors.

Parameters:

x - training samples.

y - training labels in [0, k), where k is the number of classes.

t - the inducing input, which are pre-selected or inducing samples acting as active set of regressors. Commonly, these can be chosen as the centers of k-means clustering.

Returns:

a trained Gaussian Process.