Skip to main content
Fig. 1 | Journal of Cheminformatics

Fig. 1

From: Uncertain of uncertainties? A comparison of uncertainty quantification metrics for chemical data sets

Fig. 1

a Points: a single absolute error (AE) sampled from a normal distribution with standard deviation \(\sigma\). x’s average absolute error (MAE) averaged over 100 points sampled from a normal distribution with standard deviation \(\sigma\). The black dashed line is defined by \(|\varepsilon |\) = \(\sqrt{2/\pi }\sigma\) coresponding to the MAE of a Gaussian error distribution with standard deviation \(\sigma\). The Spearman rank correlation coefficient is 0.56 and 1.0 for the dots and x’s, respectively. b Same as in (a) but in intervals of 0.01\(\sigma\). The blue and orange lines are defined by \(|\varepsilon | = 2\sigma\) and \(|\varepsilon | = 3\sigma\), respectively. The Spearman rank correlation coefficient is 0.55 and 0.99 for the dots and x’s, respectively

Back to article page