Package opennlp.tools.util.featuregen
Class WindowFeatureGenerator
java.lang.Object
opennlp.tools.util.featuregen.WindowFeatureGenerator
- All Implemented Interfaces:
AdaptiveFeatureGenerator
Generates previous (left-sided) and next (right-sided) features for a
given
AdaptiveFeatureGenerator
. The window size can be specified.
Features:
- Current token is always included unchanged,
- Previous tokens are prefixed with
p
distance, - Next tokens are prefix with
n
distance.
- See Also:
-
Field Summary
-
Constructor Summary
ConstructorDescriptionWindowFeatureGenerator
(int prevWindowSize, int nextWindowSize, AdaptiveFeatureGenerator... generators) Initializes aWindowFeatureGenerator
with the given parameters.WindowFeatureGenerator
(AdaptiveFeatureGenerator generator) Initializes aWindowFeatureGenerator
.WindowFeatureGenerator
(AdaptiveFeatureGenerator... generators) Initializes aWindowFeatureGenerator
.WindowFeatureGenerator
(AdaptiveFeatureGenerator generator, int prevWindowSize, int nextWindowSize) Initializes aWindowFeatureGenerator
with the given parameters. -
Method Summary
Modifier and TypeMethodDescriptionvoid
Informs a feature generator that the context of the adaptive data (typically a document) is no longer valid.void
createFeatures
(List<String> features, String[] tokens, int index, String[] preds) Adds the appropriate features for the token at the specifiedindex
with the specified array ofpreviousOutcomes
to the specified list of features.toString()
void
updateAdaptiveData
(String[] tokens, String[] outcomes) Informs a feature generator that the specified tokens have been classified with the corresponding set of specifiedoutcomes
.
-
Field Details
-
PREV_PREFIX
- See Also:
-
NEXT_PREFIX
- See Also:
-
-
Constructor Details
-
WindowFeatureGenerator
public WindowFeatureGenerator(AdaptiveFeatureGenerator generator, int prevWindowSize, int nextWindowSize) Initializes aWindowFeatureGenerator
with the given parameters.- Parameters:
generator
- TheAdaptiveFeatureGenerator
to apply to the window.prevWindowSize
- Size of the window to the left of the current token.nextWindowSize
- Size of the window to the right of the current token.
-
WindowFeatureGenerator
public WindowFeatureGenerator(int prevWindowSize, int nextWindowSize, AdaptiveFeatureGenerator... generators) Initializes aWindowFeatureGenerator
with the given parameters.- Parameters:
generators
- One or moreAdaptiveFeatureGenerator
to apply to the window.prevWindowSize
- Size of the window to the left of the current token.nextWindowSize
- Size of the window to the right of the current token.
-
WindowFeatureGenerator
Initializes aWindowFeatureGenerator
. The previous and next window size is 5.- Parameters:
generator
- TheAdaptiveFeatureGenerator
to apply to the window.
-
WindowFeatureGenerator
Initializes aWindowFeatureGenerator
. The previous and next window size is 5.- Parameters:
generators
- One or moreAdaptiveFeatureGenerator
to apply to the window.
-
-
Method Details
-
createFeatures
Description copied from interface:AdaptiveFeatureGenerator
Adds the appropriate features for the token at the specifiedindex
with the specified array ofpreviousOutcomes
to the specified list of features.- Specified by:
createFeatures
in interfaceAdaptiveFeatureGenerator
- Parameters:
features
- The list of features to be added to.tokens
- The tokens of the sentence or other text unit being processed.index
- The index of the token which is currently being processed.preds
- The outcomes for the tokens prior to the specified index.
-
updateAdaptiveData
Description copied from interface:AdaptiveFeatureGenerator
Informs a feature generator that the specified tokens have been classified with the corresponding set of specifiedoutcomes
.- Specified by:
updateAdaptiveData
in interfaceAdaptiveFeatureGenerator
- Parameters:
tokens
- The tokens of the sentence or other text unit which has been processed.outcomes
- The outcomes associated with the specified tokens.
-
clearAdaptiveData
public void clearAdaptiveData()Description copied from interface:AdaptiveFeatureGenerator
Informs a feature generator that the context of the adaptive data (typically a document) is no longer valid.- Specified by:
clearAdaptiveData
in interfaceAdaptiveFeatureGenerator
-
toString
-