weka.classifiers.timeseries.eval

Class TSEvalModule

java.lang.Object

weka.classifiers.timeseries.eval.TSEvalModule

Direct Known Subclasses:

ErrorModule

public abstract class TSEvalModule
extends Object

Abstract superclass of all evaluation modules.

Version:

$Revision: 45163 $

Author:

Mark Hall (mhall{[at]}pentaho{[dot]}com)

Constructor Summary

Constructors

Constructor and Description
TSEvalModule()

Method Summary

Methods

Modifier and Type	Method and Description
abstract double[]	calculateMeasure() Calculate the measure that this module represents.
abstract void	evaluateForInstance(List<weka.classifiers.evaluation.NumericPrediction> forecasts, weka.core.Instance inst) Evaluate the given forecast(s) with respect to the given test instance.
abstract String	getDefinition() Return the mathematical formula that this evaluation module computes.
abstract String	getDescription() Return the longer (single sentence) description of this evaluation module
abstract String	getEvalName() Return the short identifying name of this evaluation module
static TSEvalModule	getModule(String moduleName) Factory method for obtaining a named evaluation module.
static List<TSEvalModule>	getModuleList() Gets a list of known evaluation modules.
List<String>	getTargetFields() Get the list of target field names.
abstract void	reset() Reset the module
void	setTargetFields(List<String> targets) Set a list of target field names.
abstract String	toSummaryString() Return the summary description of the computed measure for each target.

Methods inherited from class java.lang.Object

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

Constructor Detail

TSEvalModule

public TSEvalModule()

Method Detail

getModuleList

public static List<TSEvalModule> getModuleList()

Gets a list of known evaluation modules.

Returns:

a list of known evaluation modules

getModule

public static TSEvalModule getModule(String moduleName)
                              throws IllegalArgumentException

Factory method for obtaining a named evaluation module. If the name does not match any of the known modules, then this method will assume it is fully qualified and try to instantiate it.

Parameters:

moduleName - the name of the module to obtain

Returns:

the named module

Throws:

IllegalArgumentException - if the supplied module name is unknown and can't be instantiated.

reset

public abstract void reset()

Reset the module

getEvalName

public abstract String getEvalName()

Return the short identifying name of this evaluation module

Returns:

the short identifying name of this evaluation module

getDescription

public abstract String getDescription()

Return the longer (single sentence) description of this evaluation module

Returns:

the longer description of this module

getDefinition

public abstract String getDefinition()

Return the mathematical formula that this evaluation module computes.

Returns:

the mathematical formula that this module computes.

evaluateForInstance

public abstract void evaluateForInstance(List<weka.classifiers.evaluation.NumericPrediction> forecasts,
                       weka.core.Instance inst)
                                  throws Exception

Evaluate the given forecast(s) with respect to the given test instance. Targets with missing values are ignored.

Parameters:

forecasts - a List of forecasted values. Each element corresponds to one of the targets and is assumed to be in the same order as the list of targets supplied to the setTargetFields() method.

Throws:

Exception - if the evaluation can't be completed for some reason.

calculateMeasure

public abstract double[] calculateMeasure()
                                   throws Exception

Calculate the measure that this module represents.

Returns:

the value of the measure for this module for each of the target(s).

Throws:

Exception - if the measure can't be computed for some reason.

toSummaryString

public abstract String toSummaryString()
                                throws Exception

Return the summary description of the computed measure for each target.

Returns:

the summary string description of the computed measure.

Throws:

Exception - if the measure can't be computed for some reason.

setTargetFields

public void setTargetFields(List<String> targets)

Set a list of target field names. This list must be the same as that provided to the TSForecaster via TSForecast.setFieldsToForecast()

Parameters:

targets - a List of target field names

getTargetFields

public List<String> getTargetFields()

Get the list of target field names.

Returns:

the list of target field names.