org.knime.base.node.mine.bfn.fuzzy
Class FuzzyBasisFunctionPredictorRow

java.lang.Object
  org.knime.base.node.mine.bfn.BasisFunctionPredictorRow
      org.knime.base.node.mine.bfn.fuzzy.FuzzyBasisFunctionPredictorRow

public class FuzzyBasisFunctionPredictorRow
extends BasisFunctionPredictorRow

Author:

Thomas Gabriel, University of Konstanz

Constructor Summary

	FuzzyBasisFunctionPredictorRow(ModelContentRO pp) Creates a new predictor as fuzzy rule.
protected	FuzzyBasisFunctionPredictorRow(RowKey key, DataCell classLabel, MembershipFunction[] mem, int norm) Creates a new predictor as fuzzy rule.

Method Summary

double	compose(DataRow row, double act) Composes the degree of membership by using the disjunction of the tco-norm operator.
double	computeActivation(DataRow row) Returns the compute activation of this input vector.
double	computeDistance(DataRow row)
double	computeSpread() Returns the aggregated spread of the core.
MutableDouble[]	getMaxs()
MembershipFunction	getMemship(int i) Returns the membership for one dimension.
MembershipFunction[]	getMemships()
MutableDouble[]	getMins()
int	getNorm()
int	getNrMemships() Return number of memberships which is equivalent to the number of numeric input dimensions.
int	getNrUsedFeatures()
double	overlap(BasisFunctionPredictorRow bf, boolean symmetric) Computes the overlapping of two fuzzy basisfunction based on their core spreads.
void	save(ModelContentWO pp) Saves this row into a model content.

Methods inherited from class org.knime.base.node.mine.bfn.BasisFunctionPredictorRow

getClassLabel, getDontKnowClassDegree, getId, getNumAllCoveredPattern, getNumCorrectCoveredPattern, getNumWrongCoveredPattern, getVariance, overlapping, toString

Methods inherited from class java.lang.Object

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

Constructor Detail

FuzzyBasisFunctionPredictorRow

protected FuzzyBasisFunctionPredictorRow(RowKey key,
                                         DataCell classLabel,
                                         MembershipFunction[] mem,
                                         int norm)

Creates a new predictor as fuzzy rule.

Parameters:

key - The id for this rule.

classLabel - The class label of this rule.

mem - An array of membership functions each per dimension.

norm - A fuzzy norm to combine activations via all dimensions.

FuzzyBasisFunctionPredictorRow

public FuzzyBasisFunctionPredictorRow(ModelContentRO pp)
                               throws InvalidSettingsException

Creates a new predictor as fuzzy rule.

Parameters:

pp - Content to read rule from.

Throws:

InvalidSettingsException - If the content is invalid.

Method Detail

overlap

public double overlap(BasisFunctionPredictorRow bf,
                      boolean symmetric)

Computes the overlapping of two fuzzy basisfunction based on their core spreads.

Specified by:

overlap in class BasisFunctionPredictorRow

Parameters:

bf - the other fuzzy basis function

symmetric - if the result is proportional to both basis functions, and thus symmetric, or if it is proportional to the area of the basis function on which the function is called

Returns:

a degree of overlap normalized with the overall volume of both basis functions

getNrMemships

public int getNrMemships()

Return number of memberships which is equivalent to the number of numeric input dimensions.

Returns:

Number of membership functions.

getMemship

public MembershipFunction getMemship(int i)

Returns the membership for one dimension.

Parameters:

i - Dimension index.

Returns:

A fuzzy membership function.

getMemships

public MembershipFunction[] getMemships()

Returns:

array of fuzzy membership function

save

public void save(ModelContentWO pp)

Saves this row into a model content.

Overrides:

save in class BasisFunctionPredictorRow

Parameters:

pp - the model content to save this row to

compose

public double compose(DataRow row,
                      double act)

Composes the degree of membership by using the disjunction of the tco-norm operator.

Specified by:

compose in class BasisFunctionPredictorRow

Parameters:

row - row

act - activation

Returns:

the new activation array

See Also:

computeActivation(DataRow), Norm.computeTCoNorm(double,double)

computeActivation

public double computeActivation(DataRow row)

Returns the compute activation of this input vector.

Specified by:

computeActivation in class BasisFunctionPredictorRow

Parameters:

row - input pattern

Returns:

membership degree

computeDistance

public double computeDistance(DataRow row)

Specified by:

computeDistance in class BasisFunctionPredictorRow

Parameters:

row - to compute distance with

Returns:

computes the distance between this row and the anchor

computeSpread

public double computeSpread()

Returns the aggregated spread of the core.

Specified by:

computeSpread in class BasisFunctionPredictorRow

Returns:

the overall spread of the core regions

getNrUsedFeatures

public int getNrUsedFeatures()

Specified by:

getNrUsedFeatures in class BasisFunctionPredictorRow

Returns:

number of features that have been shrunken

getNorm

public final int getNorm()

Returns:

fuzzy norm

getMins

public final MutableDouble[] getMins()

Returns:

array of minimum bounds

getMaxs

public final MutableDouble[] getMaxs()

Returns:

array of maximum bounds