F-score

https://en.wikipedia.org/wiki/F-score There are several versions of F-score: F1 score:

$F_1=2\frac{\text{precision}\cdot \text{recall}}{\text{precision}+\text{recall}}$

where:

$$\text{precision}=\frac{\text{|true positive|}}{\text{|true positive| + |false positive|}}$$ $$\text{recall}=\frac{\text{|true positive|}}{\text{|true positive| + |false negative|}}$$

$F_{\beta }$ Score

• Used when recall is considered $\beta$ times as important as precision $F_{\beta }=(1+{\beta }^2)\frac{\text{precision}\cdot \text{recall}}{({\beta }^2\cdot \text{precision})+\text{recall}}$

F0.5 Score

F_{0.5} = (1 + 0.5^2)\frac{\text{precision} \cdot \text{recall}}{(0.5^2 \cdot \text{precision})+\text{recall}}

$$-TheF0.5scoreweightsprecisionhigherthanrecall,$$