Used in statistical analysis, the F1-score (sometimes referred to just as the F-score) is a measure of machine learning model performance that combines measures of precision and accuracy, especially as used within a confusion matrix and other forms of model evaluation. The highest possible value of an F1-score is 1, indicating both perfect precision and recall. The lowest possible value of an F1-score is 0. Results usually always fall somewhere in between 0 and 1, rather than 0 or 1 themselves.