CGAL::Classification::Evaluation Class Reference

#include <CGAL/Classification/Evaluation.h>

Definition

Class to compute several measurements to evaluate the quality of a classification output.

Constructors
	Evaluation (const Label_set &labels)
	instantiates an empty evaluation object.
template<typename GroundTruthIndexRange , typename ResultIndexRange >
	Evaluation (const Label_set &labels, const GroundTruthIndexRange &ground_truth, const ResultIndexRange &result)
	instantiates an evaluation object and computes all measurements.

Modification
template<typename GroundTruthIndexRange , typename ResultIndexRange >
void	append (const GroundTruthIndexRange &ground_truth, const ResultIndexRange &result)
	appends more items to the evaluation object.

Label Evaluation
std::size_t	confusion (Label_handle ground_truth, Label_handle result)
	returns the number of items whose ground truth is ground_truth and which were classified as result.
float	precision (Label_handle label) const
	returns the precision of the training for the given label.
float	recall (Label_handle label) const
	returns the recall of the training for the given label.
float	f1_score (Label_handle label) const
	returns the \(F_1\) score of the training for the given label.
float	intersection_over_union (Label_handle label) const
	returns the intersection over union of the training for the given label.

Global Evaluation
std::size_t	number_of_misclassified_items () const
	returns the number of misclassified items.
std::size_t	number_of_items () const
	returns the total number of items used for evaluation.
float	accuracy () const
	returns the accuracy of the training.
float	mean_f1_score () const
	returns the mean \(F_1\) score of the training over all labels (see f1_score()).
float	mean_intersection_over_union () const
	returns the mean intersection over union of the training over all labels (see intersection_over_union()).

Output Formatting Functions
std::ostream &	operator<< (std::ostream &os, const Evaluation &evaluation)
	outputs the evaluation in a simple ASCII format to the stream os.
static std::ostream &	output_to_html (std::ostream &os, const Evaluation &evaluation)
	outputs the evaluation as an HTML page to the stream os.

Constructor & Destructor Documentation

Evaluation() [1/2]

CGAL::Classification::Evaluation::Evaluation

(

const Label_set &

labels

)

instantiates an empty evaluation object.

Parameters

labels

labels used.

Evaluation() [2/2]

template<typename GroundTruthIndexRange , typename ResultIndexRange >

CGAL::Classification::Evaluation::Evaluation	(	const Label_set &	labels,
		const GroundTruthIndexRange &	ground_truth,
		const ResultIndexRange &	result
	)

instantiates an evaluation object and computes all measurements.

Parameters

labels	labels used.
ground_truth	vector of label indices: it should contain the index of the corresponding label in the Label_set provided in the constructor. Input items that do not have a ground truth information should be given the value -1.
result	similar to ground_truth but contained the result of a classification.

Member Function Documentation

accuracy()

float CGAL::Classification::Evaluation::accuracy

(

)

const

returns the accuracy of the training.

Accuracy is the total number of true positives divided by the total number of provided inliers.

append()

template<typename GroundTruthIndexRange , typename ResultIndexRange >

void CGAL::Classification::Evaluation::append	(	const GroundTruthIndexRange &	ground_truth,
		const ResultIndexRange &	result
	)

appends more items to the evaluation object.

Parameters

ground_truth	vector of label indices: it should contain the index of the corresponding label in the Label_set provided in the constructor. Input items that do not have a ground truth information should be given the value -1.
result	similar to ground_truth but contained the result of a classification.

f1_score()

float CGAL::Classification::Evaluation::f1_score

(

Label_handle

label

)

const

returns the \(F_1\) score of the training for the given label.

\(F_1\) score is the harmonic mean of precision() and recall():

\[ F_1 = 2 \times \frac{precision \times recall}{precision + recall} \]

intersection_over_union()

float CGAL::Classification::Evaluation::intersection_over_union

(

Label_handle

label

)

const

returns the intersection over union of the training for the given label.

Intersection over union is the number of true positives divided by the sum of the true positives, of the false positives and of the false negatives.

precision()

float CGAL::Classification::Evaluation::precision

(

Label_handle

label

)

const

returns the precision of the training for the given label.

Precision is the number of true positives divided by the sum of the true positives and the false positives.

recall()

float CGAL::Classification::Evaluation::recall

(

Label_handle

label

)

const

returns the recall of the training for the given label.

Recall is the number of true positives divided by the sum of the true positives and the false negatives.