Fig. 4From: Large-scale prediction of activity cliffs using machine and deep learning methods of increasing complexityInfluence of data leakage on prediction accuracy. Boxplots report the distribution of AC prediction accuracy for nine ML approaches across 42 activity classes on the basis of A BA and B MCC values according to Fig. 1 in the presence (pink boxes) or absence (brown) of data leakage (i.e., compound overlap between training and test sets)Back to article page