de.dfki.madm.mlwizard.landmarking
Class AbstractLandmarker

java.lang.Object
  extended by de.dfki.madm.mlwizard.landmarking.AbstractLandmarker
Direct Known Subclasses:
AverageNodeLm, DecisionNodeLm, KNNLearnerLm, NaiveBayesLm, RandomlyChosenNodeLm, WorstNodeLm

public abstract class AbstractLandmarker
extends java.lang.Object

Abstract superclass of all landmark learner classes, used in the landmark operator. Any landmark learner as a subclass of AbstractLandmarker can be evaluated using Cross Validation depending on the value of the cross validation parameter.

Author:
Sarah Daniel

Field Summary
protected  java.lang.String name
           
protected  java.lang.String PARAMETER_PERFORMANCE_ACCURACY
           
protected  com.rapidminer.operator.performance.PerformanceVector performance
           
 
Constructor Summary
AbstractLandmarker()
           
 
Method Summary
 java.lang.Double getAccuracy()
           
 java.lang.String getName()
           
protected  java.lang.Double getPerformanceCriterion(java.lang.String criterion)
          Returns the performance criterion average value specified.
 java.util.List<java.lang.Double> getPerformanceValues()
           
 com.rapidminer.operator.performance.PerformanceVector getPerformanceVector()
           
protected  void learn(com.rapidminer.example.ExampleSet exampleSet, com.rapidminer.operator.Operator learner, java.util.Map<java.lang.String,java.lang.String> parameters)
          The Operator learner takes the ExampleSet as input and then its model is applied to the ExampleSet.
protected  void learnExampleSet(com.rapidminer.example.ExampleSet exampleSet, com.rapidminer.operator.Operator learner, java.util.Map<java.lang.String,java.lang.String> parameters)
           
protected  com.rapidminer.example.ExampleSet normalizeExampleSet(com.rapidminer.example.ExampleSet exampleSet, java.util.Map<java.lang.String,java.lang.String> parameters)
           
protected  void xValidation(com.rapidminer.example.ExampleSet exampleSet, com.rapidminer.operator.Operator learner, java.util.Map<java.lang.String,java.lang.String> parameters)
          An XValidation is applied on the Operator learner.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

PARAMETER_PERFORMANCE_ACCURACY

protected java.lang.String PARAMETER_PERFORMANCE_ACCURACY

performance

protected com.rapidminer.operator.performance.PerformanceVector performance

name

protected java.lang.String name
Constructor Detail

AbstractLandmarker

public AbstractLandmarker()
Method Detail

learnExampleSet

protected void learnExampleSet(com.rapidminer.example.ExampleSet exampleSet,
                               com.rapidminer.operator.Operator learner,
                               java.util.Map<java.lang.String,java.lang.String> parameters)
                        throws com.rapidminer.operator.OperatorCreationException,
                               com.rapidminer.operator.OperatorException
Throws:
com.rapidminer.operator.OperatorCreationException
com.rapidminer.operator.OperatorException

learn

protected void learn(com.rapidminer.example.ExampleSet exampleSet,
                     com.rapidminer.operator.Operator learner,
                     java.util.Map<java.lang.String,java.lang.String> parameters)
              throws com.rapidminer.operator.OperatorCreationException,
                     com.rapidminer.operator.OperatorException
The Operator learner takes the ExampleSet as input and then its model is applied to the ExampleSet. Finally the performance is evaluated.

Parameters:
exampleSet - the ExampleSet used as input for the Operator learner.
learner - the Operator that should be evaluated as a landmark.
parameters - the Map containing all parameters of the landmarking.
Throws:
com.rapidminer.operator.OperatorCreationException
com.rapidminer.operator.OperatorException

xValidation

protected void xValidation(com.rapidminer.example.ExampleSet exampleSet,
                           com.rapidminer.operator.Operator learner,
                           java.util.Map<java.lang.String,java.lang.String> parameters)
                    throws com.rapidminer.operator.OperatorCreationException,
                           com.rapidminer.operator.OperatorException
An XValidation is applied on the Operator learner.

Parameters:
exampleSet - the ExampleSet used as input for the XValidation process.
learner - the Operator is an inner operator of the X-Validation process that should be evaluated as a landmark.
parameters - the Map containing all parameters of the landmarking.
Throws:
com.rapidminer.operator.OperatorCreationException
com.rapidminer.operator.OperatorException

normalizeExampleSet

protected com.rapidminer.example.ExampleSet normalizeExampleSet(com.rapidminer.example.ExampleSet exampleSet,
                                                                java.util.Map<java.lang.String,java.lang.String> parameters)
                                                         throws com.rapidminer.operator.OperatorException,
                                                                com.rapidminer.operator.OperatorCreationException
Throws:
com.rapidminer.operator.OperatorException
com.rapidminer.operator.OperatorCreationException

getPerformanceCriterion

protected java.lang.Double getPerformanceCriterion(java.lang.String criterion)
Returns the performance criterion average value specified.

Parameters:
criterion - the name of the criterion to be returned.
Returns:
The makro average if it was defined and the mikro average (the current value) otherwise.

getPerformanceValues

public java.util.List<java.lang.Double> getPerformanceValues()
Returns:
a List of Double values of the performance criteria: accuracy and rootmean squared error

getAccuracy

public java.lang.Double getAccuracy()
Returns:
a Double representing the accuracy of the landmarker.

getPerformanceVector

public com.rapidminer.operator.performance.PerformanceVector getPerformanceVector()
Returns:
the PerformanceVector of the landmark.

getName

public java.lang.String getName()
Returns:
the name of the landmark.


Copyright © 2001-2009 by Rapid-I