LIBIRWLS
model Struct Reference

It represents a trained model that has been obtained using PIRWLS or PSIRWLS. More...

#include <IOStructures.h>

Data Fields

double Kgamma
 
int kernelType
 
int sparse
 
int nSVs
 
int nElem
 
double * weights
 
struct svm_sample ** x
 
double * quadratic_value
 
int maxdim
 
double bias
 
struct svm_samplefeatures
 

Detailed Description

It represents a trained model that has been obtained using PIRWLS or PSIRWLS.

This structures saves all the variables of a trained model needed to classify future data.

Field Documentation

§ bias

double model::bias

The bias term of the classification function.

§ features

struct svm_sample* model::features

Array of features.

§ kernelType

int model::kernelType

The kernel function (linear=0, rbf=1).

§ Kgamma

double model::Kgamma

Gamma parameter of the kernel function.

§ maxdim

int model::maxdim

Number of dimensions of the dataset.

§ nElem

int model::nElem

Number of features distinct than zero in the dataset.

§ nSVs

int model::nSVs

To tell if the datasets are sparse or not.

§ quadratic_value

double* model::quadratic_value

Array that contains the norm L2 of every support vector.

§ sparse

int model::sparse

To tell if the datasets are sparse or not.

§ weights

double* model::weights

The weight associated to every support vector.

§ x

struct svm_sample** model::x

The support vectors.


The documentation for this struct was generated from the following file: