Package org.tweetyproject.machinelearning

Class ClassificationTester<S extends Observation,T extends Category>

java.lang.Object

org.tweetyproject.machinelearning.ClassificationTester<S,T>

Type Parameters:

S - type of observations

T - type of categories

Direct Known Subclasses:

CrossValidator

public abstract class ClassificationTester<S extends Observation,T extends Category>
extends Object

Classes implementing this interface provide the means to test a training mechanism for performance.

The `ClassificationTester` class is responsible for testing the performance of a classifier, either by cross-validation or by directly measuring the accuracy of the classifier on a test set. The performance is measured as a value between 0 and 1, where a higher value indicates better performance.

Author:

Matthias Thimm

Constructor Summary

Constructors

Constructor	Description
ClassificationTester()	Default constructor for the `ClassificationTester` class.

Method Summary

Modifier and Type	Method	Description
double	test(Classifier classifier, TrainingSet<S,T> trainingSet)	Measures the performance of the given classifier on the given test set.
abstract double	test(Trainer<S,T> trainer, TrainingSet<S,T> trainingSet)	This method takes a trainer and a training set and returns the performance (in the range [0,1]) of the trained classifier on the training set (e.g.

Methods inherited from class java.lang.Object

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

Constructor Details

ClassificationTester

public ClassificationTester()

Default constructor for the `ClassificationTester` class.

This constructor is used to create an instance of the abstract class. Although this class cannot be instantiated directly, this constructor will be called by subclasses when they are instantiated.

Method Details

test

public abstract double test(Trainer<S,T> trainer,
 TrainingSet<S,T> trainingSet)

This method takes a trainer and a training set and returns the performance (in the range [0,1]) of the trained classifier on the training set (e.g. using cross-validation). The larger the value, the better the trained classifier.

Parameters:

trainer - some trainer

trainingSet - some training set

Returns:

the performance of the trained classifier

test

public double test(Classifier classifier,
 TrainingSet<S,T> trainingSet)

Measures the performance of the given classifier on the given test set.

Every observation from the training set is classified by the classifier and its prediction is compared with the provided category. The return value is the ratio of the correctly classified observations to the total number of observations.

Parameters:

classifier - some classifier.

trainingSet - some training set.

Returns:

the performance of the given classifier