edu.umass.cs.mallet.base.fst.confidence

Class ConstrainedForwardBackwardConfidenceEstimator

java.lang.Object

edu.umass.cs.mallet.base.fst.confidence.TransducerConfidenceEstimator

edu.umass.cs.mallet.base.fst.confidence.ConstrainedForwardBackwardConfidenceEstimator

All Implemented Interfaces:

Serializable

public class ConstrainedForwardBackwardConfidenceEstimator
extends TransducerConfidenceEstimator

Estimates the confidence of a Segment extracted by a Transducer by performing a "constrained lattice" calculation. Essentially, this sums all possible ways this segment could have been extracted and normalizes.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ConstrainedForwardBackwardConfidenceEstimator(Transducer model)

Method Summary

Modifier and Type	Method and Description
double	estimateConfidenceFor(Segment segment, Transducer.Lattice cachedLattice) Calculates the confidence in the tagging of a Segment.

Methods inherited from class edu.umass.cs.mallet.base.fst.confidence.TransducerConfidenceEstimator

estimateConfidenceFor, getSegmentConfidences, getTransducer, rankSegmentsByConfidence, rankSegmentsByConfidence

Methods inherited from class java.lang.Object

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

Constructor Detail

ConstrainedForwardBackwardConfidenceEstimator

public ConstrainedForwardBackwardConfidenceEstimator(Transducer model)

Method Detail

estimateConfidenceFor

public double estimateConfidenceFor(Segment segment,
                                    Transducer.Lattice cachedLattice)

Calculates the confidence in the tagging of a Segment.

Specified by:

estimateConfidenceFor in class TransducerConfidenceEstimator

Returns:

0-1 confidence value. higher = more confident.