eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint Struct Reference

Weighted precision and recall for a given threshold. More...

#include <eVision/EasyDeepLearning/EasyLocate/EWeightedPrecisionRecallPoint.h>

Public Member Functions
	EWeightedPrecisionRecallPoint ()
	Default constructor.
	EWeightedPrecisionRecallPoint (const eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint &other)
	Copy constructor.
eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint &	operator= (const eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint &other)
	Assignment operator.

Public Attributes
float	Threshold
	Threshold of the metric point.
float	Recall
	Weighted recall corresponding to the threshold.
float	Precision
	Weighted precision corresponding to the threshold.
float	FScore
	Weighted F-Score corresponding to the threshold.

Detailed Description

Weighted precision and recall for a given threshold.

License(s): BasicTypes

Constructor & Destructor Documentation

EWeightedPrecisionRecallPoint() [1/2]

eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::EWeightedPrecisionRecallPoint

(

)

Default constructor.

EWeightedPrecisionRecallPoint() [2/2]

eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::EWeightedPrecisionRecallPoint

(

const eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint &

other

)

Copy constructor.

Parameters

other

Weighted metric point to copy

Member Function Documentation

operator=()

eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint & eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::operator=

(

const eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint &

other

)

Assignment operator.

Parameters

other

Other metric point to assign

Member Data Documentation

FScore

float eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::FScore

Weighted F-Score corresponding to the threshold.

The weighted F-Score is the harmonic mean of the weighted recall and weighted precision.

Precision

float eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::Precision

Weighted precision corresponding to the threshold.

The weighted precision is the weighted average of the precision for each label. The precision for a label is the number of correctly predicted element divided by the number of elements predicted for that label (correctly or wrongly).

Recall

float eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::Recall

Weighted recall corresponding to the threshold.

The weighted recall is the weighted average of the recall for each label. The recall for a label is the number of correctly predicted element divided by the number of ground truth elements of that label.

Threshold

float eVision::EasyDeepLearning::EWeightedPrecisionRecallPoint::Threshold

Threshold of the metric point.