org.knime.base.node.mine.bfn.radial
Class RadialBasisFunctionFactory

java.lang.Object
  extended by org.knime.base.node.mine.bfn.BasisFunctionFactory
      extended by org.knime.base.node.mine.bfn.radial.RadialBasisFunctionFactory

public class RadialBasisFunctionFactory
extends BasisFunctionFactory

Basic interface for all basis function algorithms. Provides the function getNewBasisFunction(.) to initialise a new prototype. This interface is needed in order to create new prototypes in the general BasisFunctionLearner. Hence a BasisFunctionLearner would be initialised with an object of type BasisFunctionFactory. It is used as inter-class to init BasisFunction(s). One implementation of the BasisFunctionFactory; here represents the RadialBasisFunctionFactory object.

Author:
Thomas Gabriel, University of Konstanz
See Also:
RadialBasisFunctionLearnerRow, commit(RowKey, DataCell, DataRow)

Field Summary
(package private) static String THETA_MINUS
          Key of theta minus.
(package private) static String THETA_PLUS
          Key of theta plus.
 
Fields inherited from class org.knime.base.node.mine.bfn.BasisFunctionFactory
CLASS_COLUMN
 
Constructor Summary
protected RadialBasisFunctionFactory(double thetaMinus, double thetaPlus, int distance, DataTableSpec spec, String[] targetColumns)
          Creates a new factory for a radial basis function learner.
 
Method Summary
 BasisFunctionLearnerRow commit(RowKey key, DataCell classInfo, DataRow row)
          Creates and returns a new RadialBasisFunctionLearnerRow initialized with a center vector and a class label.
 double getThetaMinus()
          Returns the upper bound for conflicting instances.
 double getThetaPlus()
          Returns the lower bound for non-conflicting instances.
 void save(ModelContent pp)
          Saves to model content.
 
Methods inherited from class org.knime.base.node.mine.bfn.BasisFunctionFactory
createModelSpec, findDataColumns, getDistance, getMaximums, getMinimums, getModelSpec
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

THETA_MINUS

static final String THETA_MINUS
Key of theta minus.

See Also:
Constant Field Values

THETA_PLUS

static final String THETA_PLUS
Key of theta plus.

See Also:
Constant Field Values
Constructor Detail

RadialBasisFunctionFactory

protected RadialBasisFunctionFactory(double thetaMinus,
                                     double thetaPlus,
                                     int distance,
                                     DataTableSpec spec,
                                     String[] targetColumns)
Creates a new factory for a radial basis function learner.

Parameters:
thetaMinus - the upper bound activation for conflicting instances
thetaPlus - the lower bound activation for non-conflicting instances
distance - the choice of distance function
spec - the input data to learn from
targetColumns - the class info columns in the data
Method Detail

commit

public BasisFunctionLearnerRow commit(RowKey key,
                                      DataCell classInfo,
                                      DataRow row)
Creates and returns a new RadialBasisFunctionLearnerRow initialized with a center vector and a class label.

Specified by:
commit in class BasisFunctionFactory
Parameters:
key - this row's key
row - the initial center vector
classInfo - the class info
Returns:
A new basisfunction.

getThetaMinus

public final double getThetaMinus()
Returns the upper bound for conflicting instances.

Returns:
the upper bound for activation

getThetaPlus

public final double getThetaPlus()
Returns the lower bound for non-conflicting instances.

Returns:
the lower bound for activation

save

public void save(ModelContent pp)
Saves to model content.

Overrides:
save in class BasisFunctionFactory
Parameters:
pp - the model content this is saved to.


Copyright, 2003 - 2010. All rights reserved.
University of Konstanz, Germany.
Chair for Bioinformatics and Information Mining, Prof. Dr. Michael R. Berthold.
You may not modify, publish, transmit, transfer or sell, reproduce, create derivative works from, distribute, perform, display, or in any way exploit any of the content, in whole or in part, except as otherwise expressly permitted in writing by the copyright owner or as specified in the license file distributed with this product.