Confusion Matrix - True Positive, True Negative, False Positive, False Negative

AUC or ROC curve - a receiver operating characteristic curve, is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied.  Binary classification system has four possible outcomes/combination.  

Speaking in plain English,   

• True positive (TP) - equivalent (eqv.) with hit, patient is diagnosed with disease and patient has disease.
• True negative (TN) - eqv. with correct rejection, patient is diagnosed without disease and patient does not have disease.
• False positive (FP) - eqv. with false alarm, Type I error.  Patient is diagnosed with disease but in reality does not have one.
• False negative (FN) - eqv. with miss, Type II error.  Patient is diagnose without disease but in reality does have disease.

Or speaking in computer language,

Condition positive (P) - the number of real positive cases in the data
Condition negative (N) - the number of real negative cases in the data

• True Positive (TP) - if the outcome from a prediction is p and the actual value is also p.
• True Negative (TN) - if the outcome from a prediction is n and the actual value is also n. 
• False Positive (FP) -  if the outcome from a prediction is p and the actual value is n.
• False Negative (FN) - if the outcome from a prediction is n and the actual value is p. 

Or speaking in Chinese, a bit hard to understand after translation.  

		Condition
		生病 (Disease)	健康 (non Disease)
Prediction	陽性 (Positive)	TP (真陽性/a)	FP (假陽性/b)
	陰性 (Negative)	FN (假陰性/c)	TN (真陰性/d)

In a binary classifier system, the true-positive rate refers to sensitivity; and false-positive rate refers to specificity.

Sensitivity and specificity are statistical measures of the performance of a binary classification test, also known in statistics as a classification function, that are widely used in medicine:

Sensitivity (also called the true positive rate, the recall, or probability of detection in some fields) measures the proportion of actual positives that are correctly identified as such (e.g., the percentage of sick people who are correctly identified as having the condition).

Specificity (also called the true negative rate) measures the proportion of actual negatives that are correctly identified as such (e.g., the percentage of healthy people who are correctly identified as not having the condition).

Sensitivity (靈敏度/真陽性率), true positive rate (TPR), Recall, probability of detection, Power  =
sum(true positive) / sum(condition positive) = a / (a + c)

Specificity (特異度/真陰性率), Selectivity, True negative rate (TNR) =
sum(true negative) / sum (condition negative) = d / (b + d)





URL: http://doctor.get.com.tw/m/Journal/detail.aspx?no=405145
URL: https://zh.wikipedia.org/wiki/%E9%9D%88%E6%95%8F%E5%BA%A6%E5%92%8C%E7%89%B9%E7%95%B0%E5%BA%A6
URL: https://en.wikipedia.org/wiki/Receiver_operating_characteristic
URL: https://estat.pixnet.net/blog/post/61795603-roc%E6%9B%B2%E7%B7%9A-(receiver-operating-characteristic-curve)
URL: https://en.wikipedia.org/wiki/Sensitivity_and_specificity
URL: http://www.liberaldictionary.com/false-negative-2/