Package smile.base.svm

Class LASVM<T>

java.lang.Object

smile.base.svm.LASVM<T>

Type Parameters:

T - the data type of model input objects.

All Implemented Interfaces:

Serializable

public class LASVM<T>
extends Object
implements Serializable

LASVM is an approximate SVM solver that uses online approximation. It reaches accuracies similar to that of a real SVM after performing a single sequential pass through the training examples. Further benefits can be achieved using selective sampling techniques to choose which example should be considered next. LASVM requires considerably less memory than a regular SVM solver. This becomes a considerable speed advantage for large training sets.

See Also:

• Serialized Form

Constructor Summary

Constructors

Constructor	Description
LASVM(MercerKernel<T> kernel, double C, double tol)	Constructor.
LASVM(MercerKernel<T> kernel, double Cp, double Cn, double tol)	Constructor.

Method Summary

Modifier and Type	Method	Description
KernelMachine<T>	fit(T[] x, int[] y, int epochs)	Trains the model.

Methods inherited from class java.lang.Object

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

Constructor Details

LASVM

public LASVM(MercerKernel<T> kernel,
 double C,
 double tol)

Constructor.

Parameters:

kernel - the kernel.

C - the soft margin penalty parameter.

tol - the tolerance of convergence test.

LASVM

public LASVM(MercerKernel<T> kernel,
 double Cp,
 double Cn,
 double tol)

Constructor.

Parameters:

kernel - the kernel.

Cp - the soft margin penalty parameter for positive instances.

Cn - the soft margin penalty parameter for negative instances.

tol - the tolerance of convergence test.

Method Details

fit

public KernelMachine<T> fit(T[] x,
 int[] y,
 int epochs)

Trains the model.

Parameters:

x - training samples.

y - training labels.

epochs - the number of epochs, usually 1 or 2 is sufficient.

Returns:

the model.