Skip to main content
Fig. 6 | Journal of Cheminformatics

Fig. 6

From: Spectrophores as one-dimensional descriptors calculated from three-dimensional atomic properties: applications ranging from scaffold hopping to multi-target virtual screening

Fig. 6

Average recall and precision parameters calculated for a number of machine learning classification methods applied to the DUD-E datasets. Recall is defined as the ratio of the retrieved true active compounds to all active compounds in the dataset, and precision is the ratio of the retrieved true active compounds to all predicted active compounds in the dataset. Normalization parameters are indicated by the marker shapes (diamond: ‘none’; square: ‘mean’; triangle-up: ‘std’; circle: ‘all’), and treatment of stereospecificity is indicated by the marker colors (blue: ‘none’; red: ‘unique’, green: ‘all’)

Back to article page