Package weka.classifiers.trees.lmt
Class ResidualModelSelection
java.lang.Object
weka.classifiers.trees.j48.ModelSelection
weka.classifiers.trees.lmt.ResidualModelSelection
- All Implemented Interfaces:
Serializable
,RevisionHandler
Helper class for logistic model trees (weka.classifiers.trees.lmt.LMT) to implement the
splitting criterion based on residuals.
- Version:
- $Revision: 8034 $
- Author:
- Niels Landwehr
- See Also:
-
Constructor Summary
ConstructorDescriptionResidualModelSelection
(int minNumInstances) Constructor to create ResidualModelSelection object. -
Method Summary
Modifier and TypeMethodDescriptionvoid
cleanup()
Method not in useReturns the revision string.final ClassifierSplitModel
selectModel
(Instances train) Method not in usefinal ClassifierSplitModel
selectModel
(Instances data, double[][] dataZs, double[][] dataWs) Selects split based on residuals for the given dataset.final ClassifierSplitModel
selectModel
(Instances train, Instances test) Method not in use
-
Constructor Details
-
ResidualModelSelection
public ResidualModelSelection(int minNumInstances) Constructor to create ResidualModelSelection object.- Parameters:
minNumInstances
- minimum number of instances for leaves
-
-
Method Details
-
cleanup
public void cleanup()Method not in use -
selectModel
public final ClassifierSplitModel selectModel(Instances data, double[][] dataZs, double[][] dataWs) throws Exception Selects split based on residuals for the given dataset.- Throws:
Exception
-
selectModel
Method not in use- Specified by:
selectModel
in classModelSelection
-
selectModel
Method not in use- Overrides:
selectModel
in classModelSelection
-
getRevision
Returns the revision string.- Returns:
- the revision
-