Module datarush.analytics

Package com.pervasive.datarush.analytics.knn

Class ClassificationScheme

java.lang.Object

com.pervasive.datarush.analytics.knn.ClassificationScheme

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

VotingClassification

public abstract class ClassificationScheme
extends Object
implements Serializable

Determines how to classify a query record based on its nearest neighbors.

See Also:

• Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	ClassificationScheme.Classifier	Resolves the class of a query record from a set of neighbors.
static class	ClassificationScheme.Distance	Representation of the distance between two points.

Constructor Summary

Constructors

Constructor	Description
ClassificationScheme()

Method Summary

Modifier and Type	Method	Description
static ClassificationScheme	avg()	Resolves the class of the query record by doing a straight average of all neighbors.
abstract boolean	isValidTargetType(ScalarTokenType type)	Indicates whether the classification method is valid for the specified type.
abstract ClassificationScheme.Classifier	newClassifier(int k)	Creates a new classifier for use.
static ClassificationScheme	vote()	Resolves the class of the query records by simple plurality of all neighbors (alternatively, the mode of the neighbors).

Methods inherited from class java.lang.Object

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

Constructor Details

ClassificationScheme

public ClassificationScheme()

Method Details

isValidTargetType

public abstract boolean isValidTargetType(ScalarTokenType type)

Indicates whether the classification method is valid for the specified type. Some classifiers may only be applicable to continuous data.

Parameters:

type - the type of the class data

Returns:

true if the data type is compatible, false otherwise.

newClassifier

public abstract ClassificationScheme.Classifier newClassifier(int k)

Creates a new classifier for use.

Parameters:

k - the number of neighbors to expect.

Returns:

a new classifier

avg

public static ClassificationScheme avg()

Resolves the class of the query record by doing a straight average of all neighbors.

This scheme is only applicable when the target feature is a continuous data type.

Returns:

a method which averages the neighbors

vote

public static ClassificationScheme vote()

Resolves the class of the query records by simple plurality of all neighbors (alternatively, the mode of the neighbors). In case of a tie, the class of the nearest neighbor is returned.

Returns:

a method which decides by a vote of neighbors