The F-score combines sensitivity and specificity of a method, the result of which can be compared between alternative methods.
Parameters:
(1) sensitivity (S)
(2) specificity (P)
F score =
= 2 * ((sensitivity) * (specificity)) / ((sensitivity) + (specificity))
If sensitivity and specificity are approximately equal, this simplifies to:
F score =
= (shared value)
Interpretation:
• minimum F score: 0.5 (otherwise of little value)
• maximum F score: 1.0, which indicates a 100% true positive rate.
• The F score can be used to compare different methods, with preference given to the one with the highest F-score.