In statistical classification, the F-score is a measure of accuracy, based on the precision and accuracy of a classifier. It is defined by the positive real factor $\beta$, as
$$ F_\beta(x) = (1 + \beta^2) \cdot \frac{\text{precision}(x) \cdot \text{recall}(x)}{\left(\beta^2 \cdot \text{precision}(x)\right) + \text{recall}(x)}, $$
where $x$ is the class being classified.
Typically, we consider the F1-score, which is the harmonic mean between the precision and accuracy, namely
$$ F_1(x) = 2\frac{\text{precision}(x)\cdot\text{recall}(x)}{\text{precision}(x) + \text{recall}(x)}. $$