Confusion Matrix

		Test Indicator
		No	Yes
Outcome	No	a True Negative	b False Positive
	Yes	c False Negative	d True Positive
Sensitivity: $d / (c + d)$ : The proportion of observed positives that were predicted to be positive
Specificity: $a / (a + b)$ : The proportion of observed negatives that were predicted to be negatives
Changing threshold to increase Sensitivity will decrease Specificity and vice versa. The best threshold depends on the use-case.

Textbook definitions/example: "+" is disease and "-" as non-disease:
Pasted image 20240814093136.png
Pasted image 20240814093147.png

Sources: 1