Package com.imsl.datamining.supportvectormachine
package com.imsl.datamining.supportvectormachine
Support vector machines.
This package is based on the
LIBSVM version 3.1 library for support vector machines.
Class SVRegression
implements the \(\nu\)- and \(\epsilon\) formulations for
regression: \(\nu\)-SVR and \(\epsilon\)-SVR. Class
SVClassification contains
implementations of the C- and
\(\nu\) models for classification, C-SVC and
\(\nu\)-SVC. Class SVOneClass estimates the support of a
high-dimensional distribution, using the One-class SVM. All three classes are subclasses
of SupportVectorMachine. This parent class contains methods
which are available for all SVM formulations. Furthermore, all SVMs may use different kernels for
optimization and the calculation of predictions. The kernels are implemented in
classes LinearKernel,
PolynomialKernel,
RadialBasisKernel and
SigmoidKernel.-
ClassDescriptionSpecifies a data node for a support vector machine.Abstract class to specify a kernel function for support vector machines.Specifies the linear kernel for support vector machines.Specifies the polynomial kernel for support vector machines.Specifies the radial basis kernel for support vector machines.Specifies the sigmoid kernel for support vector machines.Abstract class for generating a support vector machine.Class that wraps exceptions thrown by reflective operations in core reflection.Specifies a support vector machine for classification (SVC).Class to contain model estimates after training a support vector machine.Specifies a support vector machine for the one class problem.Specifies a support vector machine for regression (SVR).