repicea.simulation.hdrelationships

Class HDRelationshipPredictor<Stand extends HDRelationshipStand,Tree extends HDRelationshipTree>

java.lang.Object

repicea.simulation.SensitivityAnalysisParameter<ModelParameterEstimates>

repicea.simulation.REpiceaPredictor

repicea.simulation.hdrelationships.HDRelationshipPredictor<Stand,Tree>

Type Parameters:

Stand - a HDRelationshipStand-derived class

Tree - a HDRelationshipTree-derived class

All Implemented Interfaces:

java.io.Serializable, HeightPredictor<Stand,Tree>

Direct Known Subclasses:

FrenchHDRelationship2014InternalPredictor, FrenchHDRelationship2018InternalPredictor

public abstract class HDRelationshipPredictor<Stand extends HDRelationshipStand,Tree extends HDRelationshipTree>
extends REpiceaPredictor
implements HeightPredictor<Stand,Tree>

The HDRelationshipPredictor class is the basic class for all HD relationships based on linear mixed-effects modelling.

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	HDRelationshipPredictor.GaussianErrorTermForHeight
protected static class	HDRelationshipPredictor.RegressionElements

Nested classes/interfaces inherited from class repicea.simulation.REpiceaPredictor

REpiceaPredictor.CruiseLine, REpiceaPredictor.ErrorTermGroup, REpiceaPredictor.IntervalNestedInPlotDefinition

Field Summary

Fields

Modifier and Type	Field and Description
protected java.util.Map<java.lang.String,java.lang.Double>	observedHeights

Fields inherited from class repicea.simulation.REpiceaPredictor

DefaultZeroIndex, isRandomEffectsVariabilityEnabled, isResidualVariabilityEnabled, listeners, oXVector

Fields inherited from class repicea.simulation.SensitivityAnalysisParameter

isParametersVariabilityEnabled

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	HDRelationshipPredictor(boolean isVariabilityEnabledEnabled) Preferred constructor.
protected	HDRelationshipPredictor(boolean isParameterVariabilityEnabled, boolean isRandomEffectVariabilityEnabled, boolean isResidualErrorVariabilityEnabled) Second constructor for greater flexibility

Method Summary

Modifier and Type	Method and Description
protected double	blupImplementation(Stand stand, HDRelationshipPredictor.RegressionElements regElement) This method accounts for the random effects in the predictions if the random effect variability is enabled.
protected abstract HDRelationshipPredictor.RegressionElements	fixedEffectsPrediction(Stand stand, Tree t, Matrix beta) This method computes the fixed effect prediction and put the prediction, the Z vector, and the species name into m_oRegressionOutput member.
protected java.lang.Enum<?>	getErrorGroup(Tree tree)
protected abstract java.util.Collection<Tree>	getTreesFromStand(Stand stand) This method selects the trees from which the blups must be calculated.
double	predictHeightM(Stand stand, Tree tree) Predicts the height for individual trees and also implements the Monte Carlo simulation automatically.
protected void	predictHeightRandomEffects(Stand stand) This method computes the best linear unbiased predictors of the random effects
protected double	residualImplementation(Tree tree, double predictedHeightWithoutResidual) This method accounts for a random deviate if the residual variability is enabled.
protected void	setSpecificResiduals(Tree tree, GaussianErrorTerm errorTerm) This method records a normalized residuals into the simulatedResidualError member which is located in the ModelBasedSimulator class.
protected boolean	wasThisTreeInitiallyMeasured(Tree tree)

Methods inherited from class repicea.simulation.REpiceaPredictor

addModelBasedSimulatorListener, doBlupsExistForThisSubject, doesThisSubjectHaveResidualErrorTerm, doRandomDeviatesExistForThisSubject, fireModelBasedSimulatorEvent, fireRandomEffectDeviateGeneratedEvent, getBlupsForThisSubject, getCruiseLineForThisSubject, getDefaultRandomEffects, getDefaultRandomEffects, getDefaultResidualError, getGaussianErrorTerms, getIntervalNestedInPlotDefinition, getParametersForThisRealization, getRandomEffectsForThisSubject, getResidualError, getResidualErrorForThisSubject, getSubjectPlusMonteCarloSpecificId, getSubjectPlusMonteCarloSpecificId, hasSubjectBeenTestedForBlups, init, recordSubjectTestedForBlups, removeModelBasedSimulatorListener, setBlupsForThisSubject, setDefaultRandomEffects, setDefaultResidualError, setDeviatesForRandomEffectsOfThisSubject, setParameterEstimates, simulateDeviatesForRandomEffectsOfThisSubject

Methods inherited from class repicea.simulation.SensitivityAnalysisParameter

getParameterEstimates

Methods inherited from class java.lang.Object

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

Field Detail

observedHeights

protected final java.util.Map<java.lang.String,java.lang.Double> observedHeights

Constructor Detail

HDRelationshipPredictor

protected HDRelationshipPredictor(boolean isVariabilityEnabledEnabled)

Preferred constructor.

Parameters:

isVariabilityEnabledEnabled - enables the variability in the parameter estimates, the random effects and the residual errors at the same time

HDRelationshipPredictor

protected HDRelationshipPredictor(boolean isParameterVariabilityEnabled,
                                  boolean isRandomEffectVariabilityEnabled,
                                  boolean isResidualErrorVariabilityEnabled)

Second constructor for greater flexibility

Parameters:

isParameterVariabilityEnabled - enables the variability in the parameter estimates

isRandomEffectVariabilityEnabled - enables the variability in the random effects

isResidualErrorVariabilityEnabled - enables the variability in the residual errors

Method Detail

predictHeightM

public double predictHeightM(Stand stand,
                             Tree tree)

Description copied from interface: HeightPredictor

Predicts the height for individual trees and also implements the Monte Carlo simulation automatically. In case of exception, it also returns -1. If the predicted height is lower than 1.3, this method returns 1.3.

Specified by:

predictHeightM in interface HeightPredictor<Stand extends HDRelationshipStand,Tree extends HDRelationshipTree>

Parameters:

stand - a HDRelationshipStand-derived instance

tree - a HDRelationshipTree-derived instance

Returns:

tree height (m)

blupImplementation

protected double blupImplementation(Stand stand,
                                    HDRelationshipPredictor.RegressionElements regElement)

This method accounts for the random effects in the predictions if the random effect variability is enabled. Otherwise, it returns 0d.

Parameters:

stand - a Stand object

regElement - a RegressionElements object

Returns:

a simulated random effect (double)

setSpecificResiduals

protected final void setSpecificResiduals(Tree tree,
                                          GaussianErrorTerm errorTerm)

This method records a normalized residuals into the simulatedResidualError member which is located in the ModelBasedSimulator class. The method asks the date from the HeightableTree instance in order to put the normalized residual at the proper location in the vector of residuals.

Parameters:

tree - a MonteCarloSimulationCompliantObject instance which stands for the tree

errorTerm - a GaussianErrorTerm instance

residualImplementation

protected double residualImplementation(Tree tree,
                                        double predictedHeightWithoutResidual)

This method accounts for a random deviate if the residual variability is enabled. Otherwise, it returns 0d.

Parameters:

tree - a HDRelationshipTree instance

Returns:

a simulated residual (double)

wasThisTreeInitiallyMeasured

protected final boolean wasThisTreeInitiallyMeasured(Tree tree)

predictHeightRandomEffects

protected void predictHeightRandomEffects(Stand stand)

This method computes the best linear unbiased predictors of the random effects

Parameters:

stand - a HeightableStand instance

getErrorGroup

protected java.lang.Enum<?> getErrorGroup(Tree tree)

getTreesFromStand

protected abstract java.util.Collection<Tree> getTreesFromStand(Stand stand)

This method selects the trees from which the blups must be calculated.

Parameters:

stand - a Stand instance

Returns:

return a Collection of Tree instances

fixedEffectsPrediction

protected abstract HDRelationshipPredictor.RegressionElements fixedEffectsPrediction(Stand stand,
                                                                                     Tree t,
                                                                                     Matrix beta)

This method computes the fixed effect prediction and put the prediction, the Z vector, and the species name into m_oRegressionOutput member. The method applies in any cases no matter it is deterministic or stochastic. NOTE: This method must be synchronized!!!!

Parameters:

stand - a Stand instance

t - a Tree instance

beta - a Matrix that contains the parameters