Classification Metrics

Counts

-		actual
		`condition_positive`	`condition_negative`
predicted	`predicted_positive`	`true_positives`	`false_positives` (`type_1_errors`)
	`predicted_negative`	`false_negatives` (`type_2_errors`)	`true_negatives`

correctly_classified vs incorrectly_classified

Fractions

Name	Synonyms	Description
`accuracy`	-	Fraction of correctly classified observations.
`f_score`	-	Accuracy measure that considers recall and precision.
`f1_score`	-	Harmonic mean of recall and precision.
`prevalence`	-	Fraction of truly positive outcomes in the data.
`positive_predictive_value`	`precision_score`	Fraction of positive predicted outcomes that are true positives.
`negative_predictive_value`	-	Fraction of negative predicted outcomes that are true negatives.
`true_positive_rate`	`recall`, `sensitivity`	Fraction of truly positive outcomes that were predicted as positive.
`false_positive_rate`	-	Fraction of truly negative outcomes that are (wrongly) predicted as positive (i.e. false positives).
`true_negative_rate`	`specificity`	Fraction of truly negative outcomes that were predicted as negative.
`false_negative_rate`	-	Fraction of truly positive outcomes that are (wrongly) predicted as negative (i.e. false negatives).
`false_discovery_rate`	-	Fraction of positive predicted outcomes that are false positives.
`false_omission_rate`	-	Fraction of negative predicted outcomes that are false negatives.
`positive_likelihood_ratio`	-	Fraction of true positive rate by false positive rate.
`negative_likelihood_ratio`	-	Fraction of false negative rate by true negative rate.
`diagnostic_odds_ratio`	-	Fraction of positive likelihood ratio by negative likelihood ratio.