cc.mallet.grmm.learning
Class PseudolikelihoodACRFTrainer.Maxable

java.lang.Object
  cc.mallet.grmm.util.CachingOptimizable.ByGradient
      cc.mallet.grmm.learning.PseudolikelihoodACRFTrainer.Maxable

All Implemented Interfaces:

Optimizable, Optimizable.ByGradientValue, java.io.Serializable

Enclosing class:

PseudolikelihoodACRFTrainer

public class PseudolikelihoodACRFTrainer.Maxable
extends CachingOptimizable.ByGradient
implements java.io.Serializable

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface cc.mallet.optimize.Optimizable

Optimizable.ByBatchGradient, Optimizable.ByCombiningBatchGradient, Optimizable.ByGISUpdate, Optimizable.ByGradient, Optimizable.ByGradientValue, Optimizable.ByHessian, Optimizable.ByValue, Optimizable.ByVotedPerceptron

Field Summary

protected java.util.BitSet

infiniteValues

Constructor Summary

protected

PseudolikelihoodACRFTrainer.Maxable(ACRF acrf, InstanceList ilist)

Method Summary

void	collectConstraints(InstanceList ilist)
protected double	computeValue()
protected void	computeValueGradient(double[] grad) Computes the gradient of the penalized log likelihood of the ACRF, and places it in cachedGradient[].
SparseVector[]	getConstraints(int cnum)
SparseVector[]	getExpectations(int cnum)
double	getGaussianPriorVariance()
int	getNumParameters()
void	getParameters(double[] buf)
void	printParameters() print weights
void	setGaussianPriorVariance(double gaussianPriorVariance)
protected void	setParametersInternal(double[] params)

Methods inherited from class cc.mallet.grmm.util.CachingOptimizable.ByGradient

getValue, getValueGradient, setCachedGradient

Methods inherited from class java.lang.Object

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

Methods inherited from interface cc.mallet.optimize.Optimizable

getParameter, setParameter, setParameters

Field Detail

infiniteValues

protected java.util.BitSet infiniteValues

Constructor Detail

PseudolikelihoodACRFTrainer.Maxable

protected PseudolikelihoodACRFTrainer.Maxable(ACRF acrf,
                                              InstanceList ilist)

Method Detail

getGaussianPriorVariance

public double getGaussianPriorVariance()

setGaussianPriorVariance

public void setGaussianPriorVariance(double gaussianPriorVariance)

getNumParameters

public int getNumParameters()

Specified by:

getNumParameters in interface Optimizable

getParameters

public void getParameters(double[] buf)

Specified by:

getParameters in interface Optimizable

setParametersInternal

protected void setParametersInternal(double[] params)

getExpectations

public SparseVector[] getExpectations(int cnum)

getConstraints

public SparseVector[] getConstraints(int cnum)

printParameters

public void printParameters()

print weights

computeValue

protected double computeValue()

Specified by:

computeValue in class CachingOptimizable.ByGradient

computeValueGradient

protected void computeValueGradient(double[] grad)

Computes the gradient of the penalized log likelihood of the ACRF, and places it in cachedGradient[]. Gradient is constraint - expectation - parameters/gaussianPriorVariance

Specified by:

computeValueGradient in class CachingOptimizable.ByGradient

collectConstraints

public void collectConstraints(InstanceList ilist)