cc.kave.repackaged.jayes.factor

Class AbstractFactor

java.lang.Object

cc.kave.repackaged.jayes.factor.AbstractFactor

All Implemented Interfaces:

Cloneable

Direct Known Subclasses:

DenseFactor, SparseFactor

public abstract class AbstractFactor
extends Object
implements Cloneable

Field Summary

Fields

Modifier and Type	Field and Description
protected Cut	cut
protected int[]	dimensionIDs
protected int[]	dimensions
protected int[]	selections
protected IArrayWrapper	values

Constructor Summary

Constructors

Constructor and Description
AbstractFactor()

Method Summary

Modifier and Type	Method and Description
AbstractFactor	clone()
abstract void	copyValues(IArrayWrapper arrayWrapper)
abstract void	fill(double d)
protected int	getDimensionFromID(int id)
int[]	getDimensionIDs()
int[]	getDimensions()
abstract int	getOverhead()
protected abstract int	getRealPosition(int virtualPosition)
double	getValue(int i)
IArrayWrapper	getValues()
boolean	isLogScale()
double[]	marginalizeAllBut(int sumDimensionID) marginalizes out all variables except the one with id sumDimensionID
void	multiplyCompatible(AbstractFactor compatible) multiply the factors.
void	multiplyCompatibleToLog(AbstractFactor factor)
void	multiplyPrepared(IArrayWrapper compatibleValues, int[] positions)
abstract int[]	prepareMultiplication(AbstractFactor compatible)
void	resetSelections()
void	select(int dimensionID, int index)
void	setDimensionIDs(int... ids) tells the Factor which variables the dimensions correspond to.
void	setDimensions(int... dimensions)
void	setLogScale(boolean isLogScale)
void	setValues(IArrayWrapper values)
void	sumPrepared(IArrayWrapper compatibleFactorValues, int[] preparedOperation)
protected void	validateCut()

Methods inherited from class java.lang.Object

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

Field Detail

dimensions

protected int[] dimensions

dimensionIDs

protected int[] dimensionIDs

values

protected IArrayWrapper values

selections

protected int[] selections

cut

protected Cut cut

Constructor Detail

AbstractFactor

public AbstractFactor()

Method Detail

copyValues

public abstract void copyValues(IArrayWrapper arrayWrapper)

prepareMultiplication

public abstract int[] prepareMultiplication(AbstractFactor compatible)

getRealPosition

protected abstract int getRealPosition(int virtualPosition)

fill

public abstract void fill(double d)

setValues

public void setValues(IArrayWrapper values)

getValues

public IArrayWrapper getValues()

getValue

public double getValue(int i)

setDimensions

public void setDimensions(int... dimensions)

getDimensions

public int[] getDimensions()

setDimensionIDs

public void setDimensionIDs(int... ids)

tells the Factor which variables the dimensions correspond to. Uniqueness and consistency of size is assumed.

Parameters:

ids - array of dimension ids

getDimensionIDs

public int[] getDimensionIDs()

getDimensionFromID

protected int getDimensionFromID(int id)

select

public void select(int dimensionID,
                   int index)

resetSelections

public void resetSelections()

setLogScale

public void setLogScale(boolean isLogScale)

isLogScale

public boolean isLogScale()

marginalizeAllBut

public double[] marginalizeAllBut(int sumDimensionID)

marginalizes out all variables except the one with id sumDimensionID

Parameters:

sumDimensionID - -1 for last dimension (default)

Returns:

marginalized variables

multiplyCompatible

public void multiplyCompatible(AbstractFactor compatible)

multiply the factors. Only compatible factors are allowed, meaning ones that have a subset of the variables of this factor (assume consistent Dimension ID / size pairs)

Parameters:

compatible - input factor

multiplyPrepared

public void multiplyPrepared(IArrayWrapper compatibleValues,
                             int[] positions)

sumPrepared

public void sumPrepared(IArrayWrapper compatibleFactorValues,
                        int[] preparedOperation)

validateCut

protected void validateCut()

clone

public AbstractFactor clone()

Overrides:

clone in class Object

multiplyCompatibleToLog

public void multiplyCompatibleToLog(AbstractFactor factor)

getOverhead

public abstract int getOverhead()

Returns:

approximated memory requirements on top of the value array