Figure 2From: A document classifier for medicinal chemistry publications trained on the ChEMBL corpusReceiver operator characteristic curve and external validation performance (Pipeline Pilot model). The ROC curve generated by a Bayesian classifier (`Learn Good From Bad’ component) in the 80% - 20% stratified partition validation is shown in (A). The performance of this classifier in the test set is shown in (B). Abbreviations: Matthews Correlation Coefficient – MCC, Receiver Operator Characteristic – ROC.Back to article page