Skip to main content
Fig. 4 | Journal of Cheminformatics

Fig. 4

From: COVER: conformational oversampling as data augmentation for molecules

Fig. 4

Sensitivity versus specificity of the trained models for different data ensembles. Each dot presents the evaluation of one model. The captions of the subplots denote the amount of oversampling done, e.g. 1–16 denotes 1 conformation for the negative class and 16 conformations for the positive class. Each color denotes an independent run of the cross-validation scheme, using a data ensemble with different conformations. The grey area indicates the area where sensitivity and specificity are greater than 0.5. a Plot of the models trained during the inner cross-validation and hyperparameter selection, b validation of the final models in the outer cross-validation loop, c validation of the final models on the external test dataset

Back to article page
\