model
Class GPRegression

java.lang.Object
  function.BasicScalarFunction
      function.VariableDomainScalarFunction
          model.BasicApproxModel
              model.GPRegression

All Implemented Interfaces:

DataListener, Computable, ScalarFunction, VariableBounds, VariableInputDimension, ApproxModel, DataFit

public class GPRegression
extends BasicApproxModel

Creation process: Set tau (compulsory), set prior mean and covariance, set bounds, init().

Author:

dgorur

Nested Class Summary

static class

GPRegression.STATE Encodes the state of the GP.

Field Summary

static double

STD_DEVS_OUT

Fields inherited from class function.BasicScalarFunction

DEFAULT_BOUNDS, DEFAULT_DIM, VALUE_OUTSIDE_BOUNDS

Constructor Summary

GPRegression()

Method Summary

void	calibrate() Maximizes log marginal likelihood of the data.
void	calibrateOptional() Optional calibration, based on some measure of calibration being required.
void	clear() Notification that all data has been removed.
double	computeLOOError() Computes the RMS Leave-One-Out error.
int	dataAdded(DataBlock block) Notification that a DataBlock has been added.
boolean	dataAdded(DataPoint dataPoint) Notification that a single DataPoint has been added.
int	dataRemoved(DataBlock dataBlock) Notification that a DataBlock has been removed.
boolean	dataRemoved(DataPoint dataPoint) Notification that a single DataPoint has been removed.
double	getCalibrationThreshold() Returns the calibrationThreshold.
double	getConsistencyThreshold() Returns the threshold at which points are considered inconsistent.
DifferentiableCorrelationFunction	getCorrelationFunction() Returns the correlation function.
BasicCorrelationFunction.TYPE	getCorrelationFunctionType() Returns the correlation function type.
java.util.HashSet<DataPoint>	getInconsistentData() Returns the set of inconsistent data.
CholeskyFactor	getL() Returns the l.
double	getMaxUncertainty()
double	getPriorMean() Returns the prior mean.
double	getPriorVariance() Returns the prior variance.
java.util.HashSet<DataPoint>	getRedundantData() Returns the set of redundant data.
GPRegression.STATE	getState() Returns the state.
double[]	getTau() Returns the length-scale parameters, tau.
double[]	getTauMax() Returns the upper bound on length-scale parameters.
double[]	getTauMin() Returns the lower bound on length-scale parameters.
cern.colt.matrix.DoubleMatrix2D	getX() Returns the x.
void	init() Initialization method.
boolean	isCalibrated() Returns true if calibrated.
boolean	isInitialized() Returns true if init() routine has been invoked.
static void	main(java.lang.String[] args)
protected ScalarFunction	makeUncertaintyEstimator()
double	quickCompute(double[] x) Compute method without dimension checking.
void	quickCompute(cern.colt.matrix.DoubleMatrix2D mat, cern.colt.matrix.DoubleMatrix1D outputs) Compute method without dimension checking, operating on DoubleMatrix2D and exploiting optimized operations therein.
void	setBounds(Hypercube bounds) Sets the bounds for this ScalarFunction.
void	setCalibrationThreshold(double calibrationThreshold) Sets the calibrationThreshold.
void	setConsistencyThreshold(double consistencyThreshold) Sets the consistency threshold.
void	setCorrelationFunctionType(BasicCorrelationFunction.TYPE correlationFunctionType) Sets the correlation function type.
void	setPriorMean(double priorMean) Sets the prior mean.
void	setPriorVariance(double priorVariance) Sets the prior variance.
void	setTau(double[] tau) Sets the length-scale parameters.
void	setTauMax(double[] tauMax) Sets the upper bound on length-scale parameters..
void	setTauMin(double[] tauMin) Sets the lower bound on length-scale parameters.

Methods inherited from class model.BasicApproxModel

getOutputNumber, getUncertaintyEstimator, setOutputNumber, setUncertaintyEstimator

Methods inherited from class function.VariableDomainScalarFunction

getBounds, getInputDimension, setInputDimension

Methods inherited from class function.BasicScalarFunction

checkDimensions, compute, compute, compute, compute, compute, compute, getOutputDimension, isWithinBounds, makeBounds

Methods inherited from class java.lang.Object

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

Methods inherited from interface function.ScalarFunction

compute, compute, compute, compute

Methods inherited from interface function.Computable

compute, compute, getBounds, getInputDimension, getOutputDimension, isWithinBounds

Methods inherited from interface function.VariableInputDimension

setInputDimension

Field Detail

STD_DEVS_OUT

public static final double STD_DEVS_OUT

Constructor Detail

GPRegression

public GPRegression()

Method Detail

main

public static void main(java.lang.String[] args)

calibrate

public void calibrate()

Maximizes log marginal likelihood of the data. See Rasmussen and Williams, Ch. 5.

calibrateOptional

public void calibrateOptional()

Description copied from interface: DataFit

Optional calibration, based on some measure of calibration being required.

clear

public void clear()

Description copied from interface: DataListener

Notification that all data has been removed.

computeLOOError

public double computeLOOError()

Computes the RMS Leave-One-Out error.

Returns:

RMS LOO error.

dataAdded

public int dataAdded(DataBlock block)

Description copied from interface: DataListener

Notification that a DataBlock has been added.

Parameters:

block - the DataBlock added.

Returns:

number of points successfully added.

dataAdded

public boolean dataAdded(DataPoint dataPoint)

Description copied from interface: DataListener

Notification that a single DataPoint has been added.

Parameters:

dataPoint - the DataPoint added.

Returns:

true if the DataPoint was successfully added.

dataRemoved

public int dataRemoved(DataBlock dataBlock)

Description copied from interface: DataListener

Notification that a DataBlock has been removed.

Parameters:

dataBlock - the DataBlock removed.

Returns:

the number of DataPoints successfully removed.

dataRemoved

public boolean dataRemoved(DataPoint dataPoint)

Description copied from interface: DataListener

Notification that a single DataPoint has been removed.

Parameters:

dataPoint - the DataPointremoved.

Returns:

true if the DataPoint was successfully removed.

getCalibrationThreshold

public double getCalibrationThreshold()

Returns the calibrationThreshold.

Returns:

the calibrationThreshold.

getConsistencyThreshold

public double getConsistencyThreshold()

Returns the threshold at which points are considered inconsistent.

Returns:

the consistencyThreshold.

getCorrelationFunction

public DifferentiableCorrelationFunction getCorrelationFunction()

Returns the correlation function.

Returns:

the correlation function.

getCorrelationFunctionType

public BasicCorrelationFunction.TYPE getCorrelationFunctionType()

Returns the correlation function type.

Returns:

the correlation function type.

getInconsistentData

public java.util.HashSet<DataPoint> getInconsistentData()

Returns the set of inconsistent data.

Returns:

the set of inconsistent data.

getL

public CholeskyFactor getL()

Returns the l.

Returns:

the l.

getMaxUncertainty

public double getMaxUncertainty()

getPriorMean

public double getPriorMean()

Returns the prior mean.

Returns:

the prior mean.

getPriorVariance

public double getPriorVariance()

Returns the prior variance.

Returns:

the prior variance.

getRedundantData

public java.util.HashSet<DataPoint> getRedundantData()

Returns the set of redundant data.

Returns:

the set of redundant data.

getState

public GPRegression.STATE getState()

Returns the state.

Returns:

the state.

getTau

public double[] getTau()

Returns the length-scale parameters, tau.

Returns:

the tau.

getTauMax

public double[] getTauMax()

Returns the upper bound on length-scale parameters.

Returns:

the upper bound on length-scale parameters.

getTauMin

public double[] getTauMin()

Returns the lower bound on length-scale parameters.

Returns:

the lower bound on length-scale parameters.

getX

public cern.colt.matrix.DoubleMatrix2D getX()

Returns the x.

Returns:

the x.

init

public void init()

Description copied from class: BasicScalarFunction

Initialization method.

Specified by:

init in interface ScalarFunction

Specified by:

init in interface DataFit

Overrides:

init in class BasicApproxModel

isCalibrated

public boolean isCalibrated()

Returns true if calibrated.

Returns:

true if calibrated.

isInitialized

public boolean isInitialized()

Returns true if init() routine has been invoked.

Returns:

true if initialized.

quickCompute

public double quickCompute(double[] x)

Description copied from class: BasicScalarFunction

Compute method without dimension checking. All subclasses must implement this method: it is the core method that defines the function.

Specified by:

quickCompute in interface ScalarFunction

Specified by:

quickCompute in class BasicScalarFunction

Parameters:

x - the given input.

Returns:

the output corresponding to the given input.

quickCompute

public void quickCompute(cern.colt.matrix.DoubleMatrix2D mat,
                         cern.colt.matrix.DoubleMatrix1D outputs)

Description copied from class: BasicScalarFunction

Compute method without dimension checking, operating on DoubleMatrix2D and exploiting optimized operations therein.

Specified by:

quickCompute in interface ScalarFunction

Overrides:

quickCompute in class BasicScalarFunction

Parameters:

mat - given matrix of inputs.

outputs - vector to store the result.

setBounds

public void setBounds(Hypercube bounds)

Description copied from interface: VariableBounds

Sets the bounds for this ScalarFunction.

Specified by:

setBounds in interface VariableBounds

Overrides:

setBounds in class VariableDomainScalarFunction

Parameters:

bounds - the given bounds.

setCalibrationThreshold

public void setCalibrationThreshold(double calibrationThreshold)

Sets the calibrationThreshold.

Parameters:

calibrationThreshold - the given calibrationThreshold.

setConsistencyThreshold

public void setConsistencyThreshold(double consistencyThreshold)

Sets the consistency threshold.

Parameters:

consistencyThreshold - the given consistency threshold.

setCorrelationFunctionType

public void setCorrelationFunctionType(BasicCorrelationFunction.TYPE correlationFunctionType)

Sets the correlation function type.

Parameters:

correlationFunctionType - the given correlation function type.

setPriorMean

public void setPriorMean(double priorMean)

Sets the prior mean.

Parameters:

priorMean - the given prior mean.

setPriorVariance

public void setPriorVariance(double priorVariance)

Sets the prior variance.

Parameters:

priorVariance - the given prior variance.

setTau

public void setTau(double[] tau)

Sets the length-scale parameters.

Parameters:

tau - the given length-scale parameters.

setTauMax

public void setTauMax(double[] tauMax)

Sets the upper bound on length-scale parameters..

Parameters:

tauMax - the given upper bound on length-scale parameters.

setTauMin

public void setTauMin(double[] tauMin)

Sets the lower bound on length-scale parameters.

Parameters:

tauMin - the given lower bound on length-scale parameters.

makeUncertaintyEstimator

protected ScalarFunction makeUncertaintyEstimator()

Specified by:

makeUncertaintyEstimator in class BasicApproxModel