@PublicationReference(author="Christopher M. Bishop",title="Pattern Recognition and Machine Learning",type=Book,year=2006,pages={152,159}) @PublicationReference(author="Hanna M. Wallach",title="Introduction to Gaussian Process Regression",type=Misc,year=2005,url="http://www.cs.umass.edu/~wallach/talks/gp_intro.pdf") @PublicationReference(author="Wikipedia",title="Bayesian linear regression",type=WebPage,year=2010,url="http://en.wikipedia.org/wiki/Bayesian_linear_regression") public class BayesianLinearRegression extends AbstractCloneableSerializable implements BayesianRegression<java.lang.Double,MultivariateGaussian>
| Modifier and Type | Class and Description |
|---|---|
static class |
BayesianLinearRegression.IncrementalEstimator
Incremental estimator for BayesianLinearRegression
|
class |
BayesianLinearRegression.PredictiveDistribution
Creates the predictive distribution for the likelihood of a given point.
|
| Modifier and Type | Field and Description |
|---|---|
static double |
DEFAULT_OUTPUT_VARIANCE
Default output variance, 1.0.
|
static double |
DEFAULT_WEIGHT_VARIANCE
Default weight variance, 1.0.
|
protected double |
outputVariance
Assumed known variance of the outputs (measurements),
must be greater than zero.
|
protected MultivariateGaussian |
weightPrior
Prior distribution of the weights, typically a zero-mean,
diagonal-variance distribution.
|
| Constructor and Description |
|---|
BayesianLinearRegression(double outputVariance,
MultivariateGaussian weightPrior)
Creates a new instance of BayesianLinearRegression
|
BayesianLinearRegression(int dimensionality)
Creates a new instance of BayesianLinearRegression
|
| Modifier and Type | Method and Description |
|---|---|
BayesianLinearRegression |
clone()
This makes public the clone method on the
Object class and
removes the exception that it throws. |
UnivariateGaussian |
createConditionalDistribution(Vectorizable input,
Vector weights)
Creates the distribution from which the outputs are generated, given
the weights and the input to consider.
|
BayesianLinearRegression.PredictiveDistribution |
createPredictiveDistribution(MultivariateGaussian posterior)
Creates the predictive distribution of outputs given the weight posterior
|
double |
getOutputVariance()
Getter for outputVariance
|
MultivariateGaussian |
getWeightPrior()
Getter for weightPrior
|
MultivariateGaussian.PDF |
learn(java.util.Collection<? extends InputOutputPair<? extends Vectorizable,java.lang.Double>> data)
The
learn method creates an object of ResultType using
data of type DataType, using some form of "learning" algorithm. |
void |
setOutputVariance(double outputVariance)
Setter for outputVariance
|
void |
setWeightPrior(MultivariateGaussian weightPrior)
Setter for weightPrior
|
public static final double DEFAULT_OUTPUT_VARIANCE
public static final double DEFAULT_WEIGHT_VARIANCE
protected double outputVariance
protected MultivariateGaussian weightPrior
public BayesianLinearRegression(int dimensionality)
dimensionality - Sets up the parameters (except featureMap) for the given dimensionality
of objects in feature space.public BayesianLinearRegression(double outputVariance,
MultivariateGaussian weightPrior)
outputVariance - Assumed known variance of the outputs (measurements),
must be greater than zero.weightPrior - Prior distribution of the weights, typically a zero-mean,
diagonal-variance distribution.public BayesianLinearRegression clone()
AbstractCloneableSerializableObject class and
removes the exception that it throws. Its default behavior is to
automatically create a clone of the exact type of object that the
clone is called on and to copy all primitives but to keep all references,
which means it is a shallow copy.
Extensions of this class may want to override this method (but call
super.clone() to implement a "smart copy". That is, to target
the most common use case for creating a copy of the object. Because of
the default behavior being a shallow copy, extending classes only need
to handle fields that need to have a deeper copy (or those that need to
be reset). Some of the methods in ObjectUtil may be helpful in
implementing a custom clone method.
Note: The contract of this method is that you must use
super.clone() as the basis for your implementation.clone in interface CloneableSerializableclone in class AbstractCloneableSerializablepublic MultivariateGaussian.PDF learn(java.util.Collection<? extends InputOutputPair<? extends Vectorizable,java.lang.Double>> data)
BatchLearnerlearn method creates an object of ResultType using
data of type DataType, using some form of "learning" algorithm.learn in interface BatchLearner<java.util.Collection<? extends InputOutputPair<? extends Vectorizable,java.lang.Double>>,MultivariateGaussian>data - The data that the learning algorithm will use to create an
object of ResultType.public UnivariateGaussian createConditionalDistribution(Vectorizable input, Vector weights)
createConditionalDistribution in interface BayesianRegression<java.lang.Double,MultivariateGaussian>input - Input to condition onweights - Weights that determine the meanpublic MultivariateGaussian getWeightPrior()
public void setWeightPrior(MultivariateGaussian weightPrior)
weightPrior - Prior distribution of the weights, typically a zero-mean,
diagonal-variance distribution.public double getOutputVariance()
public void setOutputVariance(double outputVariance)
outputVariance - Assumed known variance of the outputs (measurements),
must be greater than zero.public BayesianLinearRegression.PredictiveDistribution createPredictiveDistribution(MultivariateGaussian posterior)
createPredictiveDistribution in interface BayesianRegression<java.lang.Double,MultivariateGaussian>posterior - Posterior distribution of weights.