cc.mallet.fst.confidence
Class IsolatedSegmentTransducerCorrector

java.lang.Object
  cc.mallet.fst.confidence.IsolatedSegmentTransducerCorrector

All Implemented Interfaces:

TransducerCorrector

public class IsolatedSegmentTransducerCorrector
extends java.lang.Object
implements TransducerCorrector

Corrects a subset of the Segments produced by a Transducer. It's most useful to find the Segments that the Transducer is least confident in and correct those using the true Labeling (correctLeastConfidenceSegments). Unlike in ConstrainedViterbi, the corrected segment does not affect the labeling of other segments in the sequence. For comparison.

Constructor Summary

IsolatedSegmentTransducerCorrector(Transducer model)

IsolatedSegmentTransducerCorrector(TransducerConfidenceEstimator confidenceEstimator, Transducer model)

Method Summary

java.util.ArrayList

correctLeastConfidentSegments(InstanceList ilist, java.lang.Object[] startTags, java.lang.Object[] continueTags)

Methods inherited from class java.lang.Object

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

Constructor Detail

IsolatedSegmentTransducerCorrector

public IsolatedSegmentTransducerCorrector(TransducerConfidenceEstimator confidenceEstimator,
                                          Transducer model)

IsolatedSegmentTransducerCorrector

public IsolatedSegmentTransducerCorrector(Transducer model)

Method Detail

correctLeastConfidentSegments

public java.util.ArrayList correctLeastConfidentSegments(InstanceList ilist,
                                                         java.lang.Object[] startTags,
                                                         java.lang.Object[] continueTags)

Specified by:

correctLeastConfidentSegments in interface TransducerCorrector

Parameters:

ilist - original Transducer InstanceList

startTags - start segment tags (B-)

continueTags - continue segment tags (I-) TransducerConfidenceEstimator}

Returns:

a list of Sequences corresponding to the corrected tagging of each Instance in ilist. Note that these corrections will not affect tokens outside of the corrected segment.