---

HuberData Class Reference
[Huber]

#include <HuberData.h>

Inheritance diagram for HuberData:

List of all members.

Public Methods
	HuberData (int nobservations_in, int npredictors_in, double cutoff, double *X=0, double *y=0)
	~HuberData ()
virtual void	XtMult (double beta, SimpleVector &y, double alpha, SimpleVector &x)
virtual void	XtTransMult (double beta, SimpleVector &y, double alpha, SimpleVector &x)
virtual double	datanorm ()
virtual void	print ()
Static Public Methods
HuberData *	textInput (char filename[], double cutoff, int &iErr)
Public Attributes
int	m
int	n
DenseGenMatrixHandle	Xt
SimpleVectorHandle	Y
double	cutoff
int	nobservations
int	npredictors

---

Detailed Description

Data class for Huber.

---

Constructor & Destructor Documentation

HuberData::HuberData (  int    nobservations_in, int    npredictors_in, double    cutoff, double *    X = 0, double *    y = 0 )

make data object using data structures already allocated

HuberData::~HuberData (    )

destructor

---

Member Function Documentation

virtual double HuberData::datanorm (    )  [virtual]

compute the norm of the problem data Implements Data.

virtual void HuberData::print (    )  [inline, virtual]

print the problem data Implements Data.

HuberData* HuberData::textInput (  char    filename[], double    cutoff, int &    iErr )  [static]

reads problem data from a text file. Parameters: filename  name of input file. First entry of file is the integer "nobservations" representing number of observations. Second entry "npredictors" represents number of predictr variables. Subsequent entries represent successive rows of the predictor matrix X, with each row followed by the value of the corresponding target variable Y.

virtual void HuberData::XtMult (  double    beta, SimpleVector &    y, double    alpha, SimpleVector &    x )  [virtual]

performs operation y <- beta*y + alpha*Xt*x

virtual void HuberData::XtTransMult (  double    beta, SimpleVector &    y, double    alpha, SimpleVector &    x )  [virtual]

performs operation y <- beta*y + alpha*Xt^T*x; that is, y <- beta*y + alpha*X*x

---

Member Data Documentation

double HuberData::cutoff

cutoff parameter: value at which the loss function turns from a least-squares function into an absolute-value function

int HuberData::m

dimensions of the equivalent LCP reformulation of the Huber regression problem

int HuberData::n

dimensions of the equivalent LCP reformulation of the Huber regression problem

int HuberData::nobservations

number of observations

int HuberData::npredictors

number of predictors

DenseGenMatrixHandle HuberData::Xt

store the transpose of the matrix X as a dense matrix. Dimensions of Xt are npredictors x nobservations

SimpleVectorHandle HuberData::Y

right-hand side (targets)

---

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

• HuberData.h

---