RWRegressionModelSelector< T, S, F > Class Template Reference
[Model Selection]

The base class for the model selection classes for linear or logistic regression. More...

#include <rw/analytics/regsel.h>

List of all members.

Public Member Functions
	RWRegressionModelSelector ()
	RWRegressionModelSelector (const RWRegressionModelSelector< T, S, F > &rhs)
	RWRegressionModelSelector (const RWRegression< T, S > &reg, RWSearchMethod s)
virtual	~RWRegressionModelSelector ()
void	setRegression (const RWRegression< T, S > &r)
void	setSearchMethod (RWSearchMethod s)
const RWBitVec &	selectedParamIndices (RWSearchMethod s)
void	setCalcMethod (const RWRegressionCalc< T, S > &c)
const RWRegressionCalc< T, S > &	calcMethod () const
const RWBitVec &	selectedParamIndices () const
RWSearchMethod	searchMethod () const
bool	fail () const
RWCString	failMessage () const
const RWMathVec< T > &	selectedParamValues () const
T	evalFunctionForSelected () const
RWRegressionModelSelector< T, S, F > &	operator= (const RWRegressionModelSelector< T, S, F > &rhs)

---

Detailed Description

template<class T, class S, class F>
class RWRegressionModelSelector< T, S, F >

This is the base class for the model selection classes for linear or logistic regression. The template parameters are:

• T, the predictor variable type
• S, the observation variable type
• F, the Functor for evaluating the particular set of predictors under consideration

---

Constructor & Destructor Documentation

template<class T, class S, class F>

RWRegressionModelSelector< T, S, F >::RWRegressionModelSelector

(

)

[inline]

Default constructor. Behavior undefined.

template<class T, class S, class F>

RWRegressionModelSelector< T, S, F >::RWRegressionModelSelector

(

const RWRegressionModelSelector< T, S, F > &

rhs

)

Copy constructor.

template<class T, class S, class F>

RWRegressionModelSelector< T, S, F >::RWRegressionModelSelector	(	const RWRegression< T, S > &	reg,
		RWSearchMethod	s
	)			[inline]

Construct a selector for the given regression and search method. Once the object is constructed the search is done and it may be queried for parameter indices and so on.

template<class T, class S, class F>

virtual RWRegressionModelSelector< T, S, F >::~RWRegressionModelSelector

(

)

[inline, virtual]

Destructor.

---

Member Function Documentation

template<class T, class S, class F>

const RWRegressionCalc<T,S>& RWRegressionModelSelector< T, S, F >::calcMethod

(

)

const [inline]

Returns the parameter estimate calculation method currently in use by the model selector.

template<class T, class S, class F>

T RWRegressionModelSelector< T, S, F >::evalFunctionForSelected

(

)

const [inline]

Returns the value of the evaluation function F for the selected parameters.

template<class T, class S, class F>

bool RWRegressionModelSelector< T, S, F >::fail

(

)

const [inline]

Returns true if the search failed.

template<class T, class S, class F>

RWCString RWRegressionModelSelector< T, S, F >::failMessage

(

)

const [inline]

Returns an error message if the search fails, that is, if fail() returns true.

template<class T, class S, class F>

RWRegressionModelSelector<T,S,F>& RWRegressionModelSelector< T, S, F >::operator=

(

const RWRegressionModelSelector< T, S, F > &

rhs

)

Assignment operator.

template<class T, class S, class F>

RWSearchMethod RWRegressionModelSelector< T, S, F >::searchMethod

(

)

const [inline]

Returns the search method currently in use by the model selector.

template<class T, class S, class F>

const RWBitVec& RWRegressionModelSelector< T, S, F >::selectedParamIndices

(

)

const [inline]

Returns the indices of the parameters selected by the currently set search calculation methods.

template<class T, class S, class F>

const RWBitVec& RWRegressionModelSelector< T, S, F >::selectedParamIndices

(

RWSearchMethod

s

)

[inline]

Returns the parameters selected by the model's selector using the search method s.

template<class T, class S, class F>

const RWMathVec<T>& RWRegressionModelSelector< T, S, F >::selectedParamValues

(

)

const [inline]

Returns the currently selected parameters.

template<class T, class S, class F>

void RWRegressionModelSelector< T, S, F >::setCalcMethod

(

const RWRegressionCalc< T, S > &

c

)

Sets the calculation method used to calculate parameter estimates for the models searched.

template<class T, class S, class F>

void RWRegressionModelSelector< T, S, F >::setRegression

(

const RWRegression< T, S > &

r

)

Sets the full regression model from which the subset of predictors is selected.

template<class T, class S, class F>

void RWRegressionModelSelector< T, S, F >::setSearchMethod

(

RWSearchMethod

s

)

[inline]

Sets the search method used.