|| || || || || ||

---

Impala::Core::Training::ClassifierEvaluator Class Reference

This class works together with ParameterSearcher (through baseclass ParameterEvaluator). More...

#include <ClassifierEvaluator.h>

Inheritance diagram for Impala::Core::Training::ClassifierEvaluator:

[legend]Collaboration diagram for Impala::Core::Training::ClassifierEvaluator:

[legend]List of all members.

Public Types
typedef Table::AnnotationTable	AnnotationTable
Public Member Functions
	ClassifierEvaluator (Classifier *classifier, TrainDataSrc *src, Evaluation *evaluator)
virtual	~ClassifierEvaluator ()
void	SetRepetition (int repetition, int total)
void	SetFold (int fold, int total)
virtual double	Evaluate (Util::PropertySet *parameters)
Private Attributes
int	mRepetition
int	mFold
int	mFoldCount
Evaluation *	mEvaluator
Classifier *	mClassifier
TrainDataSrc *	mDataSrc
	ILOG_VAR_DEC

---

Detailed Description

This class works together with ParameterSearcher (through baseclass ParameterEvaluator).

In Evaluate() a classifier is trained with given parameters, this classifier is then evaluated with an instance of Evaluation. The score is returned so that the ParameterSearcher can find out what parameter setting is the best.

Currently a classifier is passed in, this would mean that classifier is retrained if training a classifier is part of the construction, then the factory will have to be passed in.

Definition at line 29 of file ClassifierEvaluator.h.

---

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

• Core/Training/ClassifierEvaluator.h

---