정말로 negative인 것 중에서 model이 positive라고 (잘못) 예측한 것의 비율.

> FPR = FP / (FP + TN)
via https://www.ibm.com/docs/en/cloud-paks/cp-data/4.6.x?topic=overview-false-positive-rate-fpr

'''False Positive Rate'''
 = (False Positives) / (Negatives)
 = 1 − (True Negatives) / (Negatives)
 = 1 - Specificity
(Kwak, Slide 1, p77)

> FPR = 1 − TNR = 1 − Specificity = FP / (FP + TN)
(chk, via https://driip.me/3ef36050-f5a3-41ea-9f23-874afe665342)

MKLINK:
[[false_positive]]는 [[type_1_error]]. - curr see [[오류,error]]
[[거짓양성,false_positive,FP]] Srch:false_positive
{
ex.
object_detection에서, 물체가 아닌데 물체로 판단
}

----
[[참양성,true_positive,TP]]
[[참음성,true_negative,TN]]
[[거짓양성,false_positive,FP]]
[[거짓음성,false_negative,FN]]

[[참양성율,true_positive_rate,TPR]] = [[민감도,sensitivity]] = [[재현율,recall]]
[[참음성율,true_negative_rate,TNR]] = [[특이도,specificity]]
'''거짓양성율,false_positive_rate,FPR'''
[[거짓음성율,false_negative_rate,FNR]]

See also
[[ROC곡선,ROC_curve]]
[[혼동행렬,confusion_matrix]]

Twins:
[[WpEn:False_positive_rate]]
... Google:false+positive+rate Naver:false+positive+rate

----
[[비율,rate]]