public class Transducer.ViterbiPath extends SequencePairAlignment
costconfidenceNBest, costNBest, input, output, outputNBest| Modifier | Constructor and Description |
|---|---|
protected |
ViterbiPath(Sequence inputSequence,
Sequence outputSequence) |
protected |
ViterbiPath(Sequence inputSequence,
Sequence outputSequence,
boolean saveLattice) |
| Modifier and Type | Method and Description |
|---|---|
Transducer.State |
getBestState(int ip) |
Sequence |
getBestStates()
Returns the maximum-value state sequence from this Lattice.
|
double |
getDelta(int ip,
int stateIndex) |
Transducer.State |
getStateAtRank(int ip,
int rank) |
Transducer |
getTransducer() |
protected edu.umass.cs.mallet.base.fst.Transducer.ViterbiPath.ViterbiNode |
getViterbiNode(edu.umass.cs.mallet.base.fst.Transducer.ViterbiPath.ViterbiNode[][] nodes,
int ip,
int stateIndex) |
void |
incrementTransducerCounts() |
double |
tokenAccuracy(Sequence referenceOutput) |
double |
tokenAccuracy(Sequence referenceOutput,
PrintWriter out) |
SequencePairAlignment |
trimStateInfo() |
getCostconfidenceNBest, costNBest, input, output, outputNBestpublic double getDelta(int ip,
int stateIndex)
public Transducer.State getBestState(int ip)
public Transducer.State getStateAtRank(int ip, int rank)
protected edu.umass.cs.mallet.base.fst.Transducer.ViterbiPath.ViterbiNode getViterbiNode(edu.umass.cs.mallet.base.fst.Transducer.ViterbiPath.ViterbiNode[][] nodes,
int ip,
int stateIndex)
public void incrementTransducerCounts()
public SequencePairAlignment trimStateInfo()
public double tokenAccuracy(Sequence referenceOutput)
public double tokenAccuracy(Sequence referenceOutput, PrintWriter out)
public Transducer getTransducer()
public Sequence getBestStates()
SequencePair.output(), which returns the Transducer's output from the
most likely state sequence.
Almost all applications will want to use the output method instead. An exception are debugging tools that care specifically about the state sequence.
Copyright © 2019 JULIE Lab, Germany. All rights reserved.