edu.stanford.nlp.ie

Class KBPStatisticalExtractor

java.lang.Object

edu.stanford.nlp.ie.KBPStatisticalExtractor

All Implemented Interfaces:

KBPRelationExtractor, java.io.Serializable

public class KBPStatisticalExtractor
extends java.lang.Object
implements KBPRelationExtractor, java.io.Serializable

A relation extractor to work with Victor's new KBP data.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface edu.stanford.nlp.ie.KBPRelationExtractor

KBPRelationExtractor.Accuracy, KBPRelationExtractor.KBPInput, KBPRelationExtractor.NERTag, KBPRelationExtractor.RelationType

Field Summary

Fields

Modifier and Type	Field and Description
Classifier<java.lang.String,java.lang.String>	classifier The implementing classifier of this extractor.
static java.lang.String	MODEL_FILE
static java.util.Optional<java.lang.String>	PREDICTIONS
static java.io.File	TEST_FILE
static java.io.File	TRAIN_FILE

Fields inherited from interface edu.stanford.nlp.ie.KBPRelationExtractor

NO_RELATION

Constructor Summary

Constructors

Constructor and Description
KBPStatisticalExtractor(Classifier<java.lang.String,java.lang.String> classifier) Create a new KBP relation extractor, from the given implementing classifier.

Method Summary

Modifier and Type	Method and Description
Pair<java.lang.String,java.lang.Double>	classify(KBPRelationExtractor.KBPInput input) Score the given input, returning both the classification decision and the probability of that decision.
static Counter<java.lang.String>	features(KBPRelationExtractor.KBPInput input)
static void	main(java.lang.String[] args)
static Classifier<java.lang.String,java.lang.String>	trainMultinomialClassifier(GeneralDataset<java.lang.String,java.lang.String> dataset, int featureThreshold, double sigma) Train a multinomial classifier off of the provided dataset.

Methods inherited from class java.lang.Object

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

Methods inherited from interface edu.stanford.nlp.ie.KBPRelationExtractor

computeAccuracy, readDataset

Field Detail

TRAIN_FILE

@ArgumentParser.Option(name="train",
                       gloss="The dataset to train on")
public static java.io.File TRAIN_FILE

TEST_FILE

@ArgumentParser.Option(name="test",
                       gloss="The dataset to test on")
public static java.io.File TEST_FILE

MODEL_FILE

@ArgumentParser.Option(name="model",
                       gloss="The dataset to test on")
public static java.lang.String MODEL_FILE

PREDICTIONS

@ArgumentParser.Option(name="predictions",
                       gloss="Dump model predictions to this file")
public static java.util.Optional<java.lang.String> PREDICTIONS

classifier

public final Classifier<java.lang.String,java.lang.String> classifier

The implementing classifier of this extractor.

Constructor Detail

KBPStatisticalExtractor

public KBPStatisticalExtractor(Classifier<java.lang.String,java.lang.String> classifier)

Create a new KBP relation extractor, from the given implementing classifier.

Parameters:

classifier - The implementing classifier.

Method Detail

features

public static Counter<java.lang.String> features(KBPRelationExtractor.KBPInput input)

trainMultinomialClassifier

public static Classifier<java.lang.String,java.lang.String> trainMultinomialClassifier(GeneralDataset<java.lang.String,java.lang.String> dataset,
                                                                                       int featureThreshold,
                                                                                       double sigma)

Train a multinomial classifier off of the provided dataset.

Parameters:

dataset - The dataset to train the classifier off of.

Returns:

A classifier.

classify

public Pair<java.lang.String,java.lang.Double> classify(KBPRelationExtractor.KBPInput input)

Score the given input, returning both the classification decision and the probability of that decision. Note that this method will not return a relation which does not type check.

Specified by:

classify in interface KBPRelationExtractor

Parameters:

input - The input to classify.

Returns:

A pair with the relation we classified into, along with its confidence.

main

public static void main(java.lang.String[] args)
                 throws java.io.IOException,
                        java.lang.ClassNotFoundException

Throws:

java.io.IOException

java.lang.ClassNotFoundException