Fig. 5
From: STAR_outliers: a python package that separates univariate outliers from non-normal distributions
STAR_outliers consistently removes closer to 0.7% of a dataset than the other algorithms (i.e. the mean absolute difference from 0.7% is closest to 0), except for [10] after setting p = 99 against its authors’ recommendations. Despite removing the correct number of points, [10] still fails to account for skew, bimodality, and monotonicity, while STAR_outliers does all of these things correctly