com.rapidminer.operator.learner.rules
Class ConjunctiveRuleModel

java.lang.Object
  com.rapidminer.operator.AbstractIOObject
      com.rapidminer.operator.ResultObjectAdapter
          com.rapidminer.operator.AbstractModel
              com.rapidminer.operator.learner.PredictionModel
                  com.rapidminer.operator.learner.SimplePredictionModel
                      com.rapidminer.operator.learner.rules.ConjunctiveRuleModel

All Implemented Interfaces:

IOObject, Model, ResultObject, Saveable, Readable, Reportable, LoggingHandler, java.io.Serializable

public class ConjunctiveRuleModel
extends SimplePredictionModel

Each object of this class represents a conjunctive rule with boolean target and nominal attributes. It cannot be changed after construction. This simplifies applications that maintain counts separately. Each attribute may be tested at most once. For two rules it can be tested whether one subsumes the other. A method for refinement allows to create each rule just once. This model may be used to query for the prediction of a single example, as well as to predict complete ExampleSets.

Author:

Martin Scholz Exp $

See Also:

Serialized Form

Constructor Summary

ConjunctiveRuleModel(ConjunctiveRuleModel ruleToExtend, Attribute attribute, double testValue)
Constructor to create an empty rule that makes a default prediction

ConjunctiveRuleModel(ConjunctiveRuleModel ruleToClone, int predictedLabel)
Constructor to clone a rule, but to change the head (prediction)

ConjunctiveRuleModel(ExampleSet exampleSet, int predictedLabel)
Constructor to create an empty rule that makes a default prediction

ConjunctiveRuleModel(ExampleSet exampleSet, int predictedLabel, int positives, int negatives)
Constructor to create an empty rule that makes a default prediction

Method Summary

boolean	equals(java.lang.Object object) Two rules are equal, if they are both permutations of the same set of literals and predict the same label.
java.util.Collection<ConjunctiveRuleModel>	getAllRefinedRules(ExampleSet exampleSet) A refinement method that - when applied sytematically during learning - generates all rules for nominal attributes and a boolean target exactly once.
Attribute	getAttributeOfLiteral(int literalNumber)
int	getConclusion()
protected int	getFirstUnusedAttribute(ExampleSet exampleSet, Attribute[] allAttributes) Helper method of getAllRefinedRules.
int	getPositionOfAttributeInRule(Attribute attribute)
int	getRuleLength()
double	getTestedValueAtLiteral(int literalNumber)
int	hashCode()
boolean	isRefinementOf(ConjunctiveRuleModel model)
double	predict(Example example) Applies the model to a single example and returns the predicted class value.
java.lang.String	toString()

Methods inherited from class com.rapidminer.operator.learner.SimplePredictionModel

performPrediction

Methods inherited from class com.rapidminer.operator.learner.PredictionModel

apply, checkCompatibility, copyPredictedLabel, createPredictedLabel, getLabel, removePredictedLabel, removePredictedLabel

Methods inherited from class com.rapidminer.operator.AbstractModel

getExtension, getFileDescription, getName, getResultIcon, getTrainingHeader, isUpdatable, setParameter, updateModel

Methods inherited from class com.rapidminer.operator.ResultObjectAdapter

addAction, getActions, getVisualizationComponent, isSavable, log, logError, logNote, logWarning, save, toHTML, toResultString

Methods inherited from class com.rapidminer.operator.AbstractIOObject

copy, getLog, getSource, initWriting, read, setLoggingHandler, setSource, write

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface com.rapidminer.operator.ResultObject

getActions, getVisualizationComponent, toResultString

Methods inherited from interface com.rapidminer.operator.IOObject

copy, getLog, getSource, setLoggingHandler, setSource, write

Methods inherited from interface com.rapidminer.operator.Saveable

isSavable, save

Constructor Detail

ConjunctiveRuleModel

public ConjunctiveRuleModel(ExampleSet exampleSet,
                            int predictedLabel)

Constructor to create an empty rule that makes a default prediction

Parameters:

exampleSet - the example set used for training

predictedLabel - specifies the head of the rule, i.e. which label to predict

ConjunctiveRuleModel

public ConjunctiveRuleModel(ExampleSet exampleSet,
                            int predictedLabel,
                            int positives,
                            int negatives)

Constructor to create an empty rule that makes a default prediction

Parameters:

exampleSet - the example set used for training

predictedLabel - specifies the head of the rule, i.e. which label to predict

ConjunctiveRuleModel

public ConjunctiveRuleModel(ConjunctiveRuleModel ruleToClone,
                            int predictedLabel)

Constructor to clone a rule, but to change the head (prediction)

ConjunctiveRuleModel

public ConjunctiveRuleModel(ConjunctiveRuleModel ruleToExtend,
                            Attribute attribute,
                            double testValue)
                     throws OperatorException

Constructor to create an empty rule that makes a default prediction

Throws:

OperatorException

Method Detail

toString

public java.lang.String toString()

Specified by:

toString in interface Readable

Overrides:

toString in class PredictionModel

Returns:

a String representation of this rule model.

predict

public double predict(Example example)
               throws OperatorException

Description copied from class: SimplePredictionModel

Applies the model to a single example and returns the predicted class value.

Specified by:

predict in class SimplePredictionModel

Throws:

OperatorException

getRuleLength

public int getRuleLength()

Returns:

the number of literals

getConclusion

public int getConclusion()

Returns:

the label this rule predicts

getAttributeOfLiteral

public Attribute getAttributeOfLiteral(int literalNumber)

Parameters:

literalNumber - the number of the literal in the rule

Returns:

the attribute tested in the specified literal

getTestedValueAtLiteral

public double getTestedValueAtLiteral(int literalNumber)

Parameters:

literalNumber - the number of the literal in the rule

Returns:

the value an attribute needs to have in order to pass the test of the specified literal

getPositionOfAttributeInRule

public int getPositionOfAttributeInRule(Attribute attribute)

Parameters:

attribute - to look for in the conjunctive rule

Returns:

the position (which is unique) of the attribute in the (ordered) conjunctive rule, or -1, if the attribute has not been found

isRefinementOf

public boolean isRefinementOf(ConjunctiveRuleModel model)

Parameters:

model - another ConjuctiveRuleModel

Returns:

true, if this rule is a refinement of the specified rule, or if both rules are equal. A rule refines another one, if it conatains all of its lietrals and predicts the same label.

equals

public boolean equals(java.lang.Object object)

Two rules are equal, if they are both permutations of the same set of literals and predict the same label.

Overrides:

equals in class java.lang.Object

hashCode

public int hashCode()

Overrides:

hashCode in class java.lang.Object

getFirstUnusedAttribute

protected int getFirstUnusedAttribute(ExampleSet exampleSet,
                                      Attribute[] allAttributes)

Helper method of getAllRefinedRules. Iterates through the Attributes of an ExampleSet and compares them to those part of the rule.

Parameters:

exampleSet -

Returns:

the index of the first attribute of the exampleSet that is not used in the rule, and for which no later Attribute is found in the rule, either.

getAllRefinedRules

public java.util.Collection<ConjunctiveRuleModel> getAllRefinedRules(ExampleSet exampleSet)
                                                              throws OperatorException

A refinement method that - when applied sytematically during learning - generates all rules for nominal attributes and a boolean target exactly once. The top-down refinement is compatible with pruning, as long as scores decrerase monotonically in support in the typical sense known from subgroup discovery. Attributes are added in the same order in which they occur in the provided exampleSet.

Parameters:

exampleSet - used to identify attributes and their values for refinement

Returns:

all refined ConjunctiveRuleModel objects

Throws:

OperatorException