magpie.attributes.evaluators

Class BaseAttributeEvaluator

java.lang.Object

magpie.attributes.evaluators.BaseAttributeEvaluator

All Implemented Interfaces:

Options

Direct Known Subclasses:

AreaUnderROCEvaluator, PearsonsCorrelationEvaluator, PolynomialFitEvaluator, SpearmansCorrelationEvaluator

public abstract class BaseAttributeEvaluator
extends java.lang.Object
implements Options

Base class for objects designed to evaluate the predictive power of single attributes. Implementing subclasses need to provide the following operations:

• evaluateAttributes_internal(magpie.data.Dataset) - Calculate predictive power of attributes
• #positiveIsBetter() - Whether a more positive parameter means better predictive ability
• #setOptions(java.lang.String[]) - Set any necessary options

Constructor Summary

Constructors

Constructor and Description
BaseAttributeEvaluator()

Method Summary

Modifier and Type	Method and Description
protected abstract java.util.Comparator<java.lang.Double>	compare() Create a comparator that will sort entries from the best to worst
protected abstract double[]	evaluateAttributes_internal(Dataset Data) Operation that actually performs the work of evaluating attributes.
double[]	evaluateAttributes(Dataset Data) Calculate the predictive power of each attributes in a dataset
int[]	getAttributeRanks(Dataset Data) Generate rank of attributes sorted by their predictive power
java.lang.String	printRankings(Dataset Data, int NumberToPrint) Print the predictive powers of the top attributes

Methods inherited from class java.lang.Object

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

Methods inherited from interface magpie.utility.interfaces.Options

printUsage, setOptions

Constructor Detail

BaseAttributeEvaluator

public BaseAttributeEvaluator()

Method Detail

evaluateAttributes

public double[] evaluateAttributes(Dataset Data)

Calculate the predictive power of each attributes in a dataset

Parameters:

Data - Dataset to evaluate

Returns:

Predictive power of each attribute

evaluateAttributes_internal

protected abstract double[] evaluateAttributes_internal(Dataset Data)

Operation that actually performs the work of evaluating attributes.

Parameters:

Data - Dataset used to evaluate attributes

Returns:

Predictive power of each attribute.

compare

protected abstract java.util.Comparator<java.lang.Double> compare()

Create a comparator that will sort entries from the best to worst

Returns:

Whether the predictive power of an attribute increases with increasing parameter

getAttributeRanks

public int[] getAttributeRanks(Dataset Data)

Generate rank of attributes sorted by their predictive power

Parameters:

Data - Dataset used to evaluate attributes

Returns:

List of attribute numbers sorted decreasingly by predictive ability (0 is best)

printRankings

public java.lang.String printRankings(Dataset Data,
                                      int NumberToPrint)

Print the predictive powers of the top attributes

Parameters:

Data - Dataset used to evaluate attributes

NumberToPrint - Number of top attributes to print

Returns:

Formatted table of attribute names, powers, and predictive capability