se.lth.cs.nlp.nlputils.ml
Class BinaryNumericClassifier<T>

java.lang.Object
  se.lth.cs.nlp.nlputils.ml.Classifier<T>
      se.lth.cs.nlp.nlputils.ml.BinaryNumericClassifier<T>

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

BinaryLinearClassifier, BinaryQuadraticClassifier

public abstract class BinaryNumericClassifier<T>
extends Classifier<T>

Abstract class for binary numeric classifiers, such as classical linear or quadratic models.

Author:

Richard Johansson (richard@cs.lth.se)

See Also:

Serialized Form

Field Summary

protected Object[]	classes The range of output values.
protected NumericEncoding	encoding The numeric encoding that maps feature values to indices.

Fields inherited from class se.lth.cs.nlp.nlputils.ml.Classifier

extractor, list

Constructor Summary

BinaryNumericClassifier()

Method Summary

double	apply(FeatureList features) Auxiliary method that encodes the features and calls apply.
abstract double	apply(int[] featureIndices) Applies the discriminator function.
Object	classify(FeatureList features) Performs a classification given a feature list.

Methods inherited from class se.lth.cs.nlp.nlputils.ml.Classifier

classify, classifyRestricted, setFeatureExtractor

Methods inherited from class java.lang.Object

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

Field Detail

classes

protected Object[] classes

The range of output values. classes[0] is returned when apply returns a positive value. Should have a length of 2.

encoding

protected NumericEncoding encoding

The numeric encoding that maps feature values to indices.

Constructor Detail

BinaryNumericClassifier

public BinaryNumericClassifier()

Method Detail

apply

public abstract double apply(int[] featureIndices)

Applies the discriminator function.

Parameters:

featureIndices - the sorted array of feature indices.

Returns:

the value of the discriminator function.

apply

public double apply(FeatureList features)

Auxiliary method that encodes the features and calls apply.

Parameters:

features - the feature list to classify.

Returns:

the value of the discriminator function.

classify

public Object classify(FeatureList features)

Description copied from class: Classifier

Performs a classification given a feature list.

Specified by:

classify in class Classifier<T>

Parameters:

features - the feature list.

Returns:

the result of the classification.