RegularizationNetwork Class Reference

Meta Model for SVM training. More...

#include <Svm.h>

Inheritance diagram for RegularizationNetwork:

MetaSVM Model

List of all members.

Public Member Functions

 RegularizationNetwork (SVM *pSVM, double gamma)
 Constructor.
 ~RegularizationNetwork ()
 Destructor.
double get_gamma ()
void set_gamma (double gamma)
bool isFeasible ()
 ensure gamma is positive

Detailed Description

Meta Model for SVM training.

Author:
T. Glasmachers
Date:
2007
The Regularization Network is a special training scheme for SVM training. It simply minimizes the regularized mean squared error resulting in a quadratic program without any constraints.

Definition at line 811 of file Svm.h.

Constructor & Destructor Documentation

RegularizationNetwork::RegularizationNetwork ( SVM *  pSVM,
double  gamma 
)

Constructor.

Definition at line 1243 of file Svm.cpp.

References Model::parameter.

RegularizationNetwork::~RegularizationNetwork (  ) 

Destructor.

Definition at line 1249 of file Svm.cpp.

Member Function Documentation

double RegularizationNetwork::get_gamma (  )  [inline]

Definition at line 821 of file Svm.h.

References Model::parameter.

Referenced by SVM_Optimizer::init().

bool RegularizationNetwork::isFeasible (  )  [virtual]

ensure gamma is positive

Reimplemented from MetaSVM.

Definition at line 1254 of file Svm.cpp.

References MetaSVM::isFeasible(), and Model::parameter.

void RegularizationNetwork::set_gamma ( double  gamma  )  [inline]

Definition at line 822 of file Svm.h.

References MetaSVM::setParameter().

The documentation for this class was generated from the following files: