StumpLearner::sRates Struct Reference

The per class rates. More...

#include <StumpLearner.h>

List of all members.

Public Member Functions

 sRates ()
 The constructor.
bool operator< (const sRates &el) const
 Overloading of the operator to allows sorting.

Public Attributes

int classIdx
 the index of the class. Needed because we will sort the vector that contains this object
double rPls
 positive rate, or $\mu_{\ell+}$
double rMin
 negative rate, or $\mu_{\ell-}$
double rZero
 abstention rate, or $\mu_{\ell0}$


Detailed Description

The per class rates.

It is defined (using $\mu$ as variable) as

\begin{eqnarray*} \mu_{\ell+} & = & \sum_{correct} w_{i,\ell}^{(t)},\\ \mu_{\ell-} &=& \sum_{incorrect} w_{i,\ell}^{(t)}, \\ \mu_{\ell0} & = & \mu_{\ell-} + \mu_{\ell+} \end{eqnarray*}

Remarks:
The implemented decision stump algorithms do not abstain themselves. The abstention value, that is for $v=0$ is computed in getEnergy()
Date:
11/11/2005

Definition at line 158 of file StumpLearner.h.


Member Function Documentation

bool operator< const sRates el  )  const [inline]
 

Overloading of the operator to allows sorting.

Parameters:
el The other element to be compared with
Date:
11/11/2005

Definition at line 173 of file StumpLearner.h.

References StumpLearner::sRates::rMin, and StumpLearner::sRates::rPls.


The documentation for this struct was generated from the following file:
Generated on Mon Nov 28 21:43:48 2005 for MultiBoost by  doxygen 1.4.5