Package opennlp.tools.namefind

Class NameFinderME

java.lang.Object

opennlp.tools.namefind.NameFinderME

All Implemented Interfaces:

TokenNameFinder

public class NameFinderME
extends Object
implements TokenNameFinder

A maximum-entropy-based name finder implementation.

Field Summary

Fields

Modifier and Type	Field	Description
static String	CONTINUE
static int	DEFAULT_BEAM_SIZE
static String	OTHER
static String	START

Constructor Summary

Constructors

Constructor	Description
NameFinderME(TokenNameFinderModel model)	Initializes a NameFinderME with a TokenNameFinderModel.

Method Summary

Modifier and Type	Method	Description
void	clearAdaptiveData()	Forgets all adaptive data which was collected during previous calls to one of the find methods.
static Span[]	dropOverlappingSpans(Span[] spans)	Removes spans with are intersecting or crossing in any way.
Span[]	find(String[] tokens)	Generates name tags for the given sequence, typically a sentence, returning token spans for any identified names.
Span[]	find(String[] tokens, String[][] additionalContext)	Generates name tags for the given sequence, typically a sentence, returning token spans for any identified names.
double[]	probs()	Retrieves the probabilities of the last decoded sequence.
void	probs(double[] probs)	Populates the specified array with the probabilities of the last decoded sequence.
double[]	probs(Span[] spans)	Retrieves an array of probabilities for each of the specified spans which is the arithmetic mean of the probabilities for each of the outcomes which make up the span.
static TokenNameFinderModel	train(String languageCode, String type, ObjectStream<NameSample> samples, TrainingParameters params, TokenNameFinderFactory factory)	Starts a training of a TokenNameFinderModel with the given parameters.

Methods inherited from class java.lang.Object

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

Field Detail

DEFAULT_BEAM_SIZE

public static final int DEFAULT_BEAM_SIZE

See Also:

Constant Field Values

START

public static final String START

See Also:

Constant Field Values

CONTINUE

public static final String CONTINUE

See Also:

Constant Field Values

OTHER

public static final String OTHER

See Also:

Constant Field Values

Constructor Detail

NameFinderME

public NameFinderME(TokenNameFinderModel model)

Initializes a NameFinderME with a TokenNameFinderModel.

Parameters:

model - The TokenNameFinderModel to initialize with.

Method Detail

find

public Span[] find(String[] tokens)

Description copied from interface: TokenNameFinder

Generates name tags for the given sequence, typically a sentence, returning token spans for any identified names.

Specified by:

find in interface TokenNameFinder

Parameters:

tokens - An array of the tokens or words of the sequence, typically a sentence.

Returns:

An array of spans for each of the names identified.

find

public Span[] find(String[] tokens,
                   String[][] additionalContext)

Generates name tags for the given sequence, typically a sentence, returning token spans for any identified names.

Parameters:

tokens - An array of the tokens or words of a sequence, typically a sentence.

additionalContext - Features which are based on context outside of the sentence but which should also be used.

Returns:

An array of token spans for each of the names identified.

clearAdaptiveData

public void clearAdaptiveData()

Description copied from interface: TokenNameFinder

Forgets all adaptive data which was collected during previous calls to one of the find methods.

Note: This method should typically be called at the end of the processing of a document.

Specified by:

clearAdaptiveData in interface TokenNameFinder

probs

public void probs(double[] probs)

Populates the specified array with the probabilities of the last decoded sequence. The sequence was determined based on the previous call to find(String[]). The specified array should be at least as large as the number of tokens in the previous call to find(String[]).

Parameters:

probs - An array with the probabilities of the last decoded sequence.

probs

public double[] probs()

Retrieves the probabilities of the last decoded sequence. The sequence was determined based on the previous call to find(String[]).

Returns:

An array with the same number of probabilities as tokens were sent to find(String[]) when it was last called.

probs

public double[] probs(Span[] spans)

Retrieves an array of probabilities for each of the specified spans which is the arithmetic mean of the probabilities for each of the outcomes which make up the span.

Parameters:

spans - The spans of the names for which probabilities are requested.

Returns:

An array of probabilities for each of the specified spans.

train

public static TokenNameFinderModel train(String languageCode,
                                         String type,
                                         ObjectStream<NameSample> samples,
                                         TrainingParameters params,
                                         TokenNameFinderFactory factory)
                                  throws IOException

Starts a training of a TokenNameFinderModel with the given parameters.

Parameters:

languageCode - The ISO conform language code.

type - The type to use.

samples - The ObjectStream of NameSample used as input for training.

params - The TrainingParameters for the context of the training.

factory - The TokenNameFinderFactory for creating related objects defined via params.

Returns:

A valid, trained TokenNameFinderModel instance.

Throws:

IOException - Thrown if IO errors occurred during training.

dropOverlappingSpans

public static Span[] dropOverlappingSpans(Span[] spans)

Removes spans with are intersecting or crossing in any way.

The following rules are used to remove the spans:
Identical spans: The first span in the array after sorting it remains.
Intersecting spans: The first span after sorting remains.
Contained spans: All spans which are contained by another are removed.

Parameters:

spans - The input spans.

Returns:

The resulting non-overlapping spans.