VariableSelectionResult Class Reference

#include <vigra/random_forest/rf_algorithm.hxx>

Public Member Functions
template<class FeatureT , class ResponseT , class ErrorRateCallBack >
bool	init (FeatureT const &all_features, ResponseT const &response, ErrorRateCallBack errorcallback)

Public Attributes
ErrorList_t	errors
double	no_features
FeatureList_t	selected

Detailed Description

Structure to hold Variable Selection results

Member Function Documentation

bool init	(	FeatureT const &	all_features,
		ResponseT const &	response,
		ErrorRateCallBack	errorcallback
	)

initialization routine. Will be called only once in the lifetime of a VariableSelectionResult. Subsequent calls will not reinitialize member variables.

This is intended, to allow continuing variable selection at a point stopped in an earlier iteration.

returns true if initialization was successful and false if the object was already initialized before.

Member Data Documentation

FeatureList_t selected

list of features.

ErrorList_t errors

vector of size (number of features)

the i-th entry encodes the error rate obtained while using features [0 - i](including i)

if the i-th entry is -1 then no error rate was obtained this may happen if more than one feature is added to the selected list in one step of the algorithm.

during initialisation error[m+n-1] is always filled

double no_features

errorrate using no features

---

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

• vigra/random_forest/rf_algorithm.hxx