|
||||||||||
PREV NEXT | FRAMES NO FRAMES |
Packages that use PredictionDistribution | |
---|---|
de.fu_berlin.ties.classify | This package provides functionality for classification of texts and feature vectors. |
de.fu_berlin.ties.classify.winnow | This package contains the Winnow classification algorithm and related algorithms and classes. |
de.fu_berlin.ties.combi | This package provides combination strategies for combining sequential classification decisions. |
de.fu_berlin.ties.filter | This packages provides generic filtering functionality and an implementation for sentence filtering. |
Uses of PredictionDistribution in de.fu_berlin.ties.classify |
---|
Methods in de.fu_berlin.ties.classify that return PredictionDistribution | |
---|---|
PredictionDistribution |
TrainableClassifier.classify(FeatureVector features,
Set candidateClasses)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
PredictionDistribution |
Classifier.classify(FeatureVector features,
Set candidateClasses)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
protected PredictionDistribution |
MetaClassifier.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
protected PredictionDistribution |
OneAgainstTheRestClassifier.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. This implementation combines the predictions for the positive class of all involved inner classifiers. |
protected PredictionDistribution |
ExternalClassifier.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
protected abstract PredictionDistribution |
TrainableClassifier.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
protected PredictionDistribution |
MultiBinaryClassifier.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. This implementation combines the predictions for the foreground of all involved inner classifiers. |
PredictionDistribution |
Reranker.rerank(PredictionDistribution origDist)
Reranks the predictions in a distribution. |
PredictionDistribution |
TrainableClassifier.trainOnError(FeatureVector features,
String targetClass,
Set candidateClasses)
Handles error-driven learning ("train only errors"): the specified feature vector is trained into the model only if the predicted class for the feature vector differs from the specified target class. |
Methods in de.fu_berlin.ties.classify with parameters of type PredictionDistribution | |
---|---|
PredictionDistribution |
Reranker.rerank(PredictionDistribution origDist)
Reranks the predictions in a distribution. |
protected boolean |
ExternalClassifier.shouldTrain(String targetClass,
PredictionDistribution predDist,
ContextMap context)
This implementation uses reinforcement training, if a thick threshold is configured. |
protected boolean |
TrainableClassifier.shouldTrain(String targetClass,
PredictionDistribution predDist,
ContextMap context)
Invoked by TrainableClassifier.trainOnError(FeatureVector, String, Set) to decide
whether to train an instance. |
protected boolean |
MetaClassifier.trainOnErrorHook(PredictionDistribution predDist,
FeatureVector features,
String targetClass,
Set candidateClasses,
ContextMap context)
Subclasses can implement this hook for more refined error-driven learning. |
protected boolean |
OneAgainstTheRestClassifier.trainOnErrorHook(PredictionDistribution predDist,
FeatureVector features,
String targetClass,
Set candidateClasses,
ContextMap context)
Subclasses can implement this hook for more refined error-driven learning. |
protected boolean |
TrainableClassifier.trainOnErrorHook(PredictionDistribution predDist,
FeatureVector features,
String targetClass,
Set candidateClasses,
ContextMap context)
Subclasses can implement this hook for more refined error-driven learning. |
protected boolean |
MultiBinaryClassifier.trainOnErrorHook(PredictionDistribution predDist,
FeatureVector features,
String targetClass,
Set candidateClasses,
ContextMap context)
Subclasses can implement this hook for more refined error-driven learning. |
Uses of PredictionDistribution in de.fu_berlin.ties.classify.winnow |
---|
Subclasses of PredictionDistribution in de.fu_berlin.ties.classify.winnow | |
---|---|
class |
WinnowDistribution
A distribution over the classes predicted by a Winnow classifier . |
Methods in de.fu_berlin.ties.classify.winnow that return PredictionDistribution | |
---|---|
protected PredictionDistribution |
Winnow.doClassify(FeatureVector features,
Set candidateClasses,
ContextMap context)
Classifies an item that is represented by a feature vector by choosing the most probable class among a set of candidate classes. |
Methods in de.fu_berlin.ties.classify.winnow with parameters of type PredictionDistribution | |
---|---|
protected boolean |
Winnow.trainOnErrorHook(PredictionDistribution predDist,
FeatureVector features,
String targetClass,
Set candidateClasses,
ContextMap context)
Hook implementing error-driven learning, promoting and demoting weights as required. |
Uses of PredictionDistribution in de.fu_berlin.ties.combi |
---|
Methods in de.fu_berlin.ties.combi with parameters of type PredictionDistribution | |
---|---|
CombinationState |
TrivialStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
CombinationState |
BeginAfterStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
CombinationState |
OpenCloseStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
CombinationState |
InsideOutsideStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
CombinationState |
BeginEndStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
abstract CombinationState |
CombinationStrategy.translateResult(PredictionDistribution[] predictions)
Translates the classes predicted by the classifiers in a decision for the caller. |
Uses of PredictionDistribution in de.fu_berlin.ties.filter |
---|
Methods in de.fu_berlin.ties.filter that return PredictionDistribution | |
---|---|
PredictionDistribution |
TrainableFilter.trainIfNecessary(Element element,
boolean decision)
Trains the correct decision for an element by calling the TrainableClassifier.trainOnError(FeatureVector, String,
java.util.Set) method on the stored trainable classifier. |
|
||||||||||
PREV NEXT | FRAMES NO FRAMES |