com.imsl.datamining.supportvectormachine

Class SigmoidKernel

java.lang.Object

com.imsl.datamining.supportvectormachine.Kernel

com.imsl.datamining.supportvectormachine.SigmoidKernel

All Implemented Interfaces:

Serializable, Cloneable

public class SigmoidKernel
extends Kernel
implements Serializable, Cloneable

Specifies the sigmoid kernel for support vector machines.

The kernel function at two data nodes, \(x_i\) and \(x_j\), is given by \(K(x_i , x_j) = \text{tanh}( \gamma x_i^T x_j + r)\) where \(\gamma\) and r are configurable parameters.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
SigmoidKernel() Constructor for the sigmoid kernel.
SigmoidKernel(double gamma, double r) Constructs a sigmoid kernel.
SigmoidKernel(SigmoidKernel sigmoidKernel) Constructs a copy of the input SigmoidKernel kernel.

Method Summary

Modifier and Type	Method and Description
SigmoidKernel	clone() Clones a SigmoidKernel kernel.
double	kernelFunction(DataNode[][] x, int i, int j) Calculates the kernel function between two DataNodes.
double	kernelFunction(DataNode[] x, DataNode[] y) Calculates the kernel function between two DataNodes.
void	setParameters(double... kParams) Sets the parameters for the sigmoid kernel.

Methods inherited from class com.imsl.datamining.supportvectormachine.Kernel

dot, getParameters

Methods inherited from class java.lang.Object

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

Constructor Detail

SigmoidKernel

public SigmoidKernel()

Constructor for the sigmoid kernel.

The default values for the sigmoid kernel are \(\gamma\)=1.0 and r=0.0.

SigmoidKernel

public SigmoidKernel(double gamma,
                     double r)

Constructs a sigmoid kernel.

Parameters:

gamma - a double, the sigmoid kernel slope parameter

r - a double, the function intercept parameter

SigmoidKernel

public SigmoidKernel(SigmoidKernel sigmoidKernel)

Constructs a copy of the input SigmoidKernel kernel.

Parameters:

sigmoidKernel - a SigmoidKernel kernel

Method Detail

clone

public SigmoidKernel clone()

Clones a SigmoidKernel kernel.

Specified by:

clone in class Kernel

Returns:

a clone of the SigmoidKernel kernel

setParameters

public void setParameters(double... kParams)

Sets the parameters for the sigmoid kernel.

Overrides:

setParameters in class Kernel

Parameters:

kParams - a double array of length 2 containing parameter values: kParams[0]=\(\gamma\) and kParams[1]=r

Default: kParams[0]=\(\gamma\)=1.0 and kParams[1]=r=0

kernelFunction

public double kernelFunction(DataNode[] x,
                             DataNode[] y)

Calculates the kernel function between two DataNodes.

Specified by:

kernelFunction in class Kernel

Parameters:

x - a DataNode array

y - a DataNode array

Returns:

a double, the kernel function evaluated at x and y

kernelFunction

public double kernelFunction(DataNode[][] x,
                             int i,
                             int j)

Calculates the kernel function between two DataNodes.

Specified by:

kernelFunction in class Kernel

Parameters:

x - a DataNode matrix

i - an int, index to the first DataNode, x[i]

j - an int, index to the second DataNode, x[j]

Returns:

a double, the kernel function evaluated at x[i] and x[j]