Figure 1

F-measure confidence intervals. 95% Confidence Intervals for the F-measure of different prediction algorithms (red circles) and the mean F-measure (black crosses). The red rectangles indicate algorithms with statistically insignificant performance differences, as determined by a permutation test.