org.openimaj.ml.linear.learner.perceptron

Class MatrixKernelPerceptron

java.lang.Object

org.openimaj.ml.linear.learner.perceptron.KernelPerceptron<double[],PerceptronClass>

org.openimaj.ml.linear.learner.perceptron.MatrixKernelPerceptron

All Implemented Interfaces:

OnlineLearner<double[],PerceptronClass>

Direct Known Subclasses:

MeanCenteredKernelPerceptron, PlusOneMatrixKernelPerceptron, Projectron, ThresholdMatrixKernelPerceptron

public class MatrixKernelPerceptron
extends KernelPerceptron<double[],PerceptronClass>

An implementation of a simple KernelPerceptron which works with Matrix inputs and is binary.

Author:

Sina Samangooei (ss@ecs.soton.ac.uk)

Field Summary

Fields

Modifier and Type	Field and Description
protected List<double[]>	supports
protected List<Double>	weights

Fields inherited from class org.openimaj.ml.linear.learner.perceptron.KernelPerceptron

errors

Constructor Summary

Constructors

Constructor and Description
MatrixKernelPerceptron(VectorKernel k)

Method Summary

Methods

Modifier and Type	Method and Description
double[]	correct(double[] in)
double	getBias()
List<double[]>	getSupports()
List<Double>	getWeights()
protected double	mapping(double[] in)
PerceptronClass	predict(double[] x)
void	update(double[] xt, PerceptronClass yt, PerceptronClass yt_prime) When there is an error in prediction, update somehow

Methods inherited from class org.openimaj.ml.linear.learner.perceptron.KernelPerceptron

getErrors, process

Methods inherited from class java.lang.Object

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

Field Detail

supports

protected List<double[]> supports

weights

protected List<Double> weights

Constructor Detail

MatrixKernelPerceptron

public MatrixKernelPerceptron(VectorKernel k)

Parameters:

k - the kernel

Method Detail

correct

public double[] correct(double[] in)

mapping

protected double mapping(double[] in)

predict

public PerceptronClass predict(double[] x)

update

public void update(double[] xt,
          PerceptronClass yt,
          PerceptronClass yt_prime)

Description copied from class: KernelPerceptron

When there is an error in prediction, update somehow

Specified by:

update in class KernelPerceptron<double[],PerceptronClass>

getSupports

public List<double[]> getSupports()

Specified by:

getSupports in class KernelPerceptron<double[],PerceptronClass>

Returns:

the vectors that form the support

getWeights

public List<Double> getWeights()

Specified by:

getWeights in class KernelPerceptron<double[],PerceptronClass>

Returns:

the weights of the support vectors

getBias

public double getBias()

Specified by:

getBias in class KernelPerceptron<double[],PerceptronClass>

Returns:

the bias