edu.udo.cs.yale.operator.features.construction
Class GeneratingForwardSelection

java.lang.Object
  edu.udo.cs.yale.operator.Operator
      edu.udo.cs.yale.operator.OperatorChain
          edu.udo.cs.yale.operator.features.FeatureOperator
              edu.udo.cs.yale.operator.features.selection.FeatureSelectionOperator
                  edu.udo.cs.yale.operator.features.construction.GeneratingForwardSelection

All Implemented Interfaces:

ConfigurationListener

public class GeneratingForwardSelection
extends FeatureSelectionOperator

This operator is a kind of nested forward selection and thus is (in contrast to a genetic algorithm) a directed search.

1. use forward selection in order to determine the best attributes
2. Create a new attribute by multiplying any of the original attributes with any of the attributes selected by the forward selection in the last turn
3. loop as long as performance increases

Version:

$Id: GeneratingForwardSelection.java,v 1.1 2006/04/14 11:42:27 ingomierswa Exp $

Author:

Simon Fischer, Ingo Mierswa

Field Summary

private Individual	bestIndividual
private int	newAttributeStart
private Attribute[]	originalAttributes List of AttributeReferences.
private int	turn
private java.util.List<FeatureGenerator>	useGenerators

Fields inherited from class edu.udo.cs.yale.operator.features.selection.FeatureSelectionOperator

BACKWARD_ELIMINATION, FORWARD_SELECTION

Constructor Summary

GeneratingForwardSelection(OperatorDescription description)

Method Summary

IOObject[]	apply() Applies the feature operator: collects the pre- and postevaluation operators create an initial population evaluate the initial population loop as long as solution is not good enough apply all pre evaluation operators evaluate the population update the population's best individual apply all post evaluation operators return all generation's best individual
Population	createInitialPopulation(ExampleSet es) May es have n features.
java.util.List<ParameterType>	getParameterTypes() Returns a list of ParameterTypes describing the parameters of this operator.
java.util.List<PopulationOperator>	getPreEvaluationPopulationOperators(ExampleSet input) The operators performs two steps: forward selection/backward elimination kick out all but the keep_best individuals remove redundant individuals
boolean	solutionGoodEnough(Population pop) Returns true if the best individual is not better than the last generation's best individual.

Methods inherited from class edu.udo.cs.yale.operator.features.selection.FeatureSelectionOperator

getPostEvaluationPopulationOperators

Methods inherited from class edu.udo.cs.yale.operator.features.FeatureOperator

evaluate, evaluate, getCheckForMaximum, getInnerOperatorCondition, getInputClasses, getMaxNumberOfInnerOperators, getMinNumberOfInnerOperators, getNumberOfSteps, getOutputClasses, getRandom, setCheckForMaximum

Methods inherited from class edu.udo.cs.yale.operator.OperatorChain

addAddListener, addOperator, addOperator, checkDeprecations, checkIO, checkNumberOfInnerOperators, checkProperties, clearErrorList, clearStepCounter, cloneOperator, countStep, createExperimentTree, delete, experimentFinished, experimentStarts, getAllInnerOperators, getCurrentStep, getIndexOfOperator, getInnerOperatorForName, getInnerOperatorsXML, getNumberOfAllOperators, getNumberOfChildrensSteps, getNumberOfOperators, getOperator, getOperatorFromAll, getOperators, isEnabled, performAdditionalChecks, removeAddListener, removeOperator, setEnabled, setExperiment, shouldReturnInnerOutput

Methods inherited from class edu.udo.cs.yale.operator.Operator

addError, addValue, addWarning, apply, createExperimentTree, createFromXML, createMarkedExperimentTree, getAddOnlyAdditionalOutput, getApplyCount, getDeliveredOutputClasses, getDeprecationInfo, getDesiredInputClasses, getErrorList, getExperiment, getInput, getInput, getInput, getInputDescription, getIOContainerForInApplyLoopBreakpoint, getName, getOperatorClassName, getOperatorDescription, getParameter, getParameterAsBoolean, getParameterAsColor, getParameterAsDouble, getParameterAsFile, getParameterAsInt, getParameterAsString, getParameterList, getParameters, getParameterType, getParent, getStartTime, getStatus, getUserDescription, getValue, getValues, getXML, hasBreakpoint, hasBreakpoint, hasInput, inApplyLoop, isParameterSet, logMessage, register, remove, rename, resume, setBreakpoint, setInput, setListParameter, setOperatorParameters, setParameter, setParameters, setParent, setUserDescription, toString, writeXML

Methods inherited from class java.lang.Object

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

Field Detail

originalAttributes

private Attribute[] originalAttributes

List of AttributeReferences.

bestIndividual

private Individual bestIndividual

useGenerators

private java.util.List<FeatureGenerator> useGenerators

newAttributeStart

private int newAttributeStart

turn

private int turn

Constructor Detail

GeneratingForwardSelection

public GeneratingForwardSelection(OperatorDescription description)

Method Detail

apply

public IOObject[] apply()
                 throws OperatorException

Description copied from class: FeatureOperator

Applies the feature operator:

1. collects the pre- and postevaluation operators
2. create an initial population
3. evaluate the initial population
4. loop as long as solution is not good enough
   1. apply all pre evaluation operators
   2. evaluate the population
   3. update the population's best individual
   4. apply all post evaluation operators
5. return all generation's best individual

Overrides:

apply in class FeatureSelectionOperator

Returns:

the last inner operator's output or the input itself if the chain is empty.

Throws:

OperatorException

createInitialPopulation

public Population createInitialPopulation(ExampleSet es)

May es have n features. The initial population contains

n elements with exactly 1 feature switched on.

Overrides:

createInitialPopulation in class FeatureSelectionOperator

getPreEvaluationPopulationOperators

public java.util.List<PopulationOperator> getPreEvaluationPopulationOperators(ExampleSet input)
                                                                       throws OperatorException

The operators performs two steps:

1. forward selection/backward elimination
2. kick out all but the keep_best individuals
3. remove redundant individuals

Overrides:

getPreEvaluationPopulationOperators in class FeatureSelectionOperator

Throws:

OperatorException

solutionGoodEnough

public boolean solutionGoodEnough(Population pop)
                           throws OperatorException

Description copied from class: FeatureSelectionOperator

Returns true if the best individual is not better than the last generation's best individual.

Overrides:

solutionGoodEnough in class FeatureSelectionOperator

Throws:

OperatorException

getParameterTypes

public java.util.List<ParameterType> getParameterTypes()

Description copied from class: Operator

Returns a list of ParameterTypes describing the parameters of this operator. The default implementation returns an empty list if no input objects can be retained and special parameters for those input objects which can be prevented from being consumed.

Overrides:

getParameterTypes in class FeatureSelectionOperator