Distributions

Table of contents

show_confusion_matrix()
show_true_positive_distribution()
- Models comparison
show_false_positive_distribution()
- Models comparison
show_false_negative_distribution()
- Models comparison
show_true_negative_distribution()
- Models comparison
show_true_positive_distribution_for_categories_for_property()
- Models comparison
show_false_positive_distribution_for_categories_for_property()
- Models comparison
show_false_negative_distribution_for_categories_for_property()
- Models comparison
show_true_negative_distribution_for_categories_for_property()
- Models comparison

show_confusion_matrix()

It shows the confusion matrix of the model. The confusion matrix can be performed for the entire data set or for a subset with a specific property value.

Parameters

categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
properties_names: list, optional; List of properties to be included in the analysis. If not specified, all the properties are included.
(default is None)
properties_values: list of list, optional; Properties values to be considered in te analysis. If not specified, all the values are considered.
(default is None)
The index of the properties values must be the same of the properties names. Example:
properties_names=['name1', 'name2']
properties_values=[['value1_of_name1', 'value2_of_name1'], ['value1_of_name2']]
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_confusion_matrix()

analyze_cm_output_a analyze_cm_output_b

N.B. As example, it is shown only the output of a single category, but the analysis is performed for all the categories selected. The confusion matrix among the categories is supported only for single-label classification task.

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
yes	yes	yes	no	no

show_true_positive_distribution()

It provides the true positive distribution among the categories.

Parameters

categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_true_positive_distribution()

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_true_positive_distribution()

show_true_positive_distribution_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_true_positive_distribution()

Localization

from odin.classes import ComparatorLocalization

my_comparator = ComparatorLocalization(dataset_gt_param, task_type, models_proposals)
my_comparator.show_true_positive_distribution()

comparator_tp_distribution

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
no	yes	yes	yes	yes

show_false_positive_distribution()

It provides the false positive distribution among the categories.

Parameters

categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_false_positive_distribution()

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_false_positive_distribution()

show_false_positive_distribution_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_false_positive_distribution()

Localization

from odin.classes import ComparatorLocalization

my_comparator = ComparatorLocalization(dataset_gt_param, task_type, models_proposals)
my_comparator.show_false_positive_distribution()

comparator_fp_distribution

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
no	yes	yes	yes	yes

show_false_negative_distribution()

It provides the false negative distribution among the categories.

Parameters

categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_false_negative_distribution()

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_false_negative_distribution()

show_false_negative_distribution_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_false_negative_distribution()

Localization

from odin.classes import ComparatorLocalization

my_comparator = ComparatorLocalization(dataset_gt_param, task_type, models_proposals)
my_comparator.show_false_negative_distribution()

comparator_fn_distribution

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
no	yes	yes	yes	yes

show_true_negative_distribution()

It provides the true negative distribution among the categories.

Parameters

categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_true_negative_distribution()

show_true_negative_distribution_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_true_negative_distribution()

comparator_tn_distribution

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
no	yes	yes	no	no

show_true_positive_distribution_for_categories_for_property()

It provides the true positive distribution of the property values for each category.

Parameters

property_name: str; Name of the property to be analyzed.
property_values: list, optional; List of the property values to be included in the analysis. If not specified, all the values are included.
(default is None)
categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_true_positive_distribution_for_categories_for_property("property_name")

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_true_positive_distribution_for_categories_for_property("property_name")

show_true_positive_distribution_for_categories_for_property_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_true_positive_distribution_for_categories_for_property("property_name")

Localization

from odin.classes import ComparatorLocalization

my_comparator = ComparatorLocalization(dataset_gt_param, task_type, models_proposals)
my_comparator.show_true_positive_distribution_for_categories_for_property("property_name")

comparator_show_true_positive_distribution_for_categories_for_property_output

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
yes	yes	yes	yes	yes

show_false_positive_distribution_for_categories_for_property()

It provides the false positive distribution of the property values for each category.

Parameters

property_name: str; Name of the property to be analyzed.
property_values: list, optional; List of the property values to be included in the analysis. If not specified, all the values are included.
(default is None)
categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_false_positive_distribution_for_categories_for_property("property_name")

show_false_positive_distribution_for_categories_for_property_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_false_positive_distribution_for_categories_for_property("property_name")

comparator_show_false_positive_distribution_for_categories_for_property_output

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
yes	yes	yes	no	no

show_false_negative_distribution_for_categories_for_property()

It provides the false negative distribution of the property values for each category.

Parameters

property_name: str; Name of the property to be analyzed.
property_values: list, optional; List of the property values to be included in the analysis. If not specified, all the values are included.
(default is None)
categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_false_negative_distribution_for_categories_for_property("property_name")

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_false_negative_distribution_for_categories_for_property("property_name")

show_false_negative_distribution_for_categories_for_property_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_false_negative_distribution_for_categories_for_property("property_name")

Localization

from odin.classes import ComparatorLocalization

my_comparator = ComparatorLocalization(dataset_gt_param, task_type, models_proposals)
my_comparator.show_false_negative_distribution_for_categories_for_property("property_name")

comparator_show_false_negative_distribution_for_categories_for_property_output

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
yes	yes	yes	yes	yes

show_true_negative_distribution_for_categories_for_property()

It provides the true negative distribution of the property values for each category.

Parameters

property_name: str; Name of the property to be analyzed.
property_values: list, optional; List of the property values to be included in the analysis. If not specified, all the values are included.
(default is None)
categories: list, optional; List of categories to be included in the analysis. If not specified, all the categories are included.
(default is None)
show: bool, optional; Indicates whether the plot should be shown or not. If False, returns the results as dict.
(default is True)

Example

Classification

from odin.classes import AnalyzerClassification

my_analyzer = AnalyzerClassification("my_classifier_name", my_classification_dataset)
my_analyzer.show_true_negative_distribution_for_categories_for_property("property_name")

Localization

from odin.classes import AnalyzerLocalization

my_analyzer = AnalyzerLocalization("my_detector_name", my_localization_dataset)
my_analyzer.show_true_negative_distribution_for_categories_for_property("property_name")

show_true_negative_distribution_for_categories_for_property_output

Models comparison

Example

Classification

from odin.classes import ComparatorClassification

my_comparator = ComparatorClassification(dataset_gt_param, classification_type, models_proposals)
my_comparator.show_true_negative_distribution_for_categories_for_property("property_name")

comparator_show_true_negative_distribution_for_categories_for_property_output

Tasks supported

Binary Classification	Single-label Classification	Multi-label Classification	Object Detection	Instance Segmentation
yes	yes	yes	no	no