edu.cmu.minorthird.classify.sequential
Class StackedSequenceLearner

java.lang.Object
  edu.cmu.minorthird.classify.sequential.StackedSequenceLearner

All Implemented Interfaces:

BatchSequenceClassifierLearner, SequenceClassifierLearner

public class StackedSequenceLearner
extends java.lang.Object
implements BatchSequenceClassifierLearner

Author:

William Cohen

Nested Class Summary

static class

StackedSequenceLearner.StackingParams Bundle of parameters for the StackedSequenceLearner.

Constructor Summary

StackedSequenceLearner()

StackedSequenceLearner(ClassifierLearner baseLearner, int depth)

StackedSequenceLearner(ClassifierLearner baseLearner, int depth, int windowSize)

StackedSequenceLearner(SequenceClassifierLearner baseLearner, int depth)

StackedSequenceLearner(SequenceClassifierLearner baseLearner, int depth, int windowSize)

Method Summary

SequenceClassifier	batchTrain(SequenceDataset dataset)
int	getHistorySize() Number of instances before the current target for which the predicted class will be used as a feature.
StackedSequenceLearner.StackingParams	getParams()
void	setHistorySize(int newHistorySize)
void	setSchema(ExampleSchema schema)
SequenceDataset	stackDataset(SequenceDataset dataset) Create a new dataset in which each instance has been augmented with the history features constructed from the *predicted* labels of previous examples, where the prediction is made using cross-validation.

Methods inherited from class java.lang.Object

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

Constructor Detail

StackedSequenceLearner

public StackedSequenceLearner()

StackedSequenceLearner

public StackedSequenceLearner(SequenceClassifierLearner baseLearner,
                              int depth)

StackedSequenceLearner

public StackedSequenceLearner(ClassifierLearner baseLearner,
                              int depth)

StackedSequenceLearner

public StackedSequenceLearner(SequenceClassifierLearner baseLearner,
                              int depth,
                              int windowSize)

StackedSequenceLearner

public StackedSequenceLearner(ClassifierLearner baseLearner,
                              int depth,
                              int windowSize)

Method Detail

getHistorySize

public int getHistorySize()

Number of instances before the current target for which the predicted class will be used as a feature.

Specified by:

getHistorySize in interface SequenceClassifierLearner

setHistorySize

public void setHistorySize(int newHistorySize)

getParams

public StackedSequenceLearner.StackingParams getParams()

setSchema

public void setSchema(ExampleSchema schema)

Specified by:

setSchema in interface SequenceClassifierLearner

batchTrain

public SequenceClassifier batchTrain(SequenceDataset dataset)

Specified by:

batchTrain in interface BatchSequenceClassifierLearner

stackDataset

public SequenceDataset stackDataset(SequenceDataset dataset)

Create a new dataset in which each instance has been augmented with the history features constructed from the *predicted* labels of previous examples, where the prediction is made using cross-validation.