it.uniroma2.sag.kelp.predictionfunction.model

Class BinaryKernelMachineModel

java.lang.Object

it.uniroma2.sag.kelp.predictionfunction.model.BinaryModel

it.uniroma2.sag.kelp.predictionfunction.model.BinaryKernelMachineModel

All Implemented Interfaces:

KernelMachineModel, Model

public class BinaryKernelMachineModel
extends BinaryModel
implements KernelMachineModel

It is the model for a binary kernel machine consisting of an implicit hyperplane in the Reproducing Kernel Hilbert Space.

Author:

Simone Filice

Field Summary

Fields inherited from class it.uniroma2.sag.kelp.predictionfunction.model.BinaryModel

bias

Constructor Summary

Constructors

Constructor and Description
BinaryKernelMachineModel()
BinaryKernelMachineModel(Kernel kernel)

Method Summary

Methods

Modifier and Type	Method and Description
void	addExample(float weight, Example example) Adds an example to the model with a given weight.
void	addSupportVector(SupportVector supportVector) Adds a support vector NOTE: it does not check whether a support vector with the same instance of the given supportVector is already in the model
Kernel	getKernel()
int	getNumberOfSupportVectors()
float	getSquaredNorm() Computes the squared norm of the hyperplane this model is based on
float	getSquaredNorm(Example example) Computes the squared norm of a given example according to the space in which the model is operating
SupportVector	getSupportVector(Example instance) Returns the support vector associated to a given instance, null the instance is not a support vector in this model
Integer	getSupportVectorIndex(Example instance) Returns the index of the vector associated to a given instance, null the instance is not a support vector in this model
List<SupportVector>	getSupportVectors() Returns all the support vectors
boolean	isSupportVector(Example instance) Returns whether instance is a support vector in this model
void	reset() Resets the model parameters to the default state.
void	setKernel(Kernel kernel)
void	setSupportVector(SupportVector sv, int position) Substitutes the support vector in position position with sv
void	setSupportVectors(List<SupportVector> supportVectors)
void	substituteSupportVector(int index, Example newInstance, float newWeight)

Methods inherited from class it.uniroma2.sag.kelp.predictionfunction.model.BinaryModel

getBias, setBias

Methods inherited from class java.lang.Object

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

Constructor Detail

BinaryKernelMachineModel

public BinaryKernelMachineModel(Kernel kernel)

BinaryKernelMachineModel

public BinaryKernelMachineModel()

Method Detail

getKernel

public Kernel getKernel()

Specified by:

getKernel in interface KernelMachineModel

Returns:

the kernel

setKernel

public void setKernel(Kernel kernel)

Specified by:

setKernel in interface KernelMachineModel

Parameters:

kernel - the kernel to set

getSupportVectors

public List<SupportVector> getSupportVectors()

Returns all the support vectors

Returns:

all the support vectors

setSupportVectors

public void setSupportVectors(List<SupportVector> supportVectors)

Parameters:

supportVectors - the supportVectors to set

addSupportVector

public void addSupportVector(SupportVector supportVector)

Adds a support vector NOTE: it does not check whether a support vector with the same instance of the given supportVector is already in the model

Parameters:

supportVector - the new support vector to be added

reset

public void reset()

Description copied from interface: Model

Resets the model parameters to the default state. All the learned stuff is forgotten

Specified by:

reset in interface Model

addExample

public void addExample(float weight,
              Example example)

Description copied from class: BinaryModel

Adds an example to the model with a given weight. The model will become a linear combination of the current model and the example

Specified by:

addExample in class BinaryModel

Parameters:

weight - the weight of the new example to add

example - the example to add

getSquaredNorm

public float getSquaredNorm(Example example)

Description copied from class: BinaryModel

Computes the squared norm of a given example according to the space in which the model is operating

Specified by:

getSquaredNorm in class BinaryModel

Parameters:

example - the example whose squared norm is required

Returns:

the squared norm of a given example

getSupportVector

public SupportVector getSupportVector(Example instance)

Returns the support vector associated to a given instance, null the instance is not a support vector in this model

Parameters:

instance - the instance whose corresponding support vector must be retrieved

Returns:

the support vector corresponding to instance

getSupportVectorIndex

public Integer getSupportVectorIndex(Example instance)

Returns the index of the vector associated to a given instance, null the instance is not a support vector in this model

Parameters:

instance - the instance whose corresponding support vector must be retrieved

Returns:

the index of the support vector corresponding to instance

isSupportVector

public boolean isSupportVector(Example instance)

Description copied from interface: KernelMachineModel

Returns whether instance is a support vector in this model

Specified by:

isSupportVector in interface KernelMachineModel

Parameters:

instance - the instance

Returns:

whether instance is a support vector in this model

setSupportVector

public void setSupportVector(SupportVector sv,
                    int position)

Substitutes the support vector in position position with sv

Parameters:

sv - the new support vector to be added

position - the position in which sv must be added

substituteSupportVector

public void substituteSupportVector(int index,
                           Example newInstance,
                           float newWeight)

getNumberOfSupportVectors

public int getNumberOfSupportVectors()

Specified by:

getNumberOfSupportVectors in interface KernelMachineModel

getSquaredNorm

public float getSquaredNorm()

Description copied from class: BinaryModel

Computes the squared norm of the hyperplane this model is based on

Specified by:

getSquaredNorm in class BinaryModel

Returns:

the squared norm of the hyperplane