jsat.classifiers

Interface UpdateableClassifier

All Superinterfaces:

Classifier, Cloneable, Serializable

All Known Implementing Classes:

ALMA2, ALMA2K, AMM, AODE, AROW, BaseUpdateableClassifier, BOGD, CSKLR, DUOL, Forgetron, KernelSGD, LinearSGD, MultinomialNaiveBayes, NaiveBayesUpdateable, NHERD, ODE, OnlineAMM, OSKL, PassiveAggressive, Projectron, ROMMA, SCW, SPA, STGD, UpdatableStacking

public interface UpdateableClassifier
extends Classifier

UpdateableClassifier is an interface for one type of Online learner. The main characteristic of an online learning is that new example points can be added incremental after the classifier was initially trained, or as part of its initial training.
Some Online learners behave differently in when they are updated. The UpdateableClassifier is an online learner that specifically only performs additional learning when a new example is provided via the update(jsat.classifiers.DataPoint, int) method.
The standard behavior for an Updateable Classifier is that the user first calls Classifier.trainC(jsat.classifiers.ClassificationDataSet) to first train the classifier, or setUp(jsat.classifiers.CategoricalData[], int, jsat.classifiers.CategoricalData) to prepare for online updates. Once one of these is called, it should then be safe to call update(jsat.classifiers.DataPoint, int) without getting a UntrainedModelException. Some online learners may require one of the train methods to be called first.

Author:

Edward Raff

Method Summary

Modifier and Type	Method and Description
UpdateableClassifier	clone()
void	setUp(CategoricalData[] categoricalAttributes, int numericAttributes, CategoricalData predicting) Prepares the classifier to begin learning from its update(jsat.classifiers.DataPoint, int) method.
void	update(DataPoint dataPoint, int targetClass) Updates the classifier by giving it a new data point to learn from.

Methods inherited from interface jsat.classifiers.Classifier

classify, supportsWeightedData, trainC, trainC

Method Detail

setUp

void setUp(CategoricalData[] categoricalAttributes,
           int numericAttributes,
           CategoricalData predicting)

Prepares the classifier to begin learning from its update(jsat.classifiers.DataPoint, int) method.

Parameters:

categoricalAttributes - an array containing the categorical attributes that will be in each data point

numericAttributes - the number of numeric attributes that will be in each data point

predicting - the information for the target class that will be predicted

update

void update(DataPoint dataPoint,
            int targetClass)

Updates the classifier by giving it a new data point to learn from.

Parameters:

dataPoint - the data point to learn

targetClass - the target class of the data point

clone

UpdateableClassifier clone()

Specified by:

clone in interface Classifier