Skip to main content
Fig. 7 | Journal of Cheminformatics

Fig. 7

From: GraphDTI: A robust deep learning predictor of drug-target interactions from multiple heterogeneous data

Fig. 7

Optimization of the number of clusters for cross-validation. Silhouette coefficient values are calculated for the GraphDTI dataset partitioned with the k-medoids algorithm into a varying number of clusters. Three measures of distances between drug-protein pairs are used in the dataset clustering, the Feature Match Distance (FMD, blue), the Perfect Match Distance (PMD, red), and the scaled PMD (green). For a given number of clusters a mean value (circles) with the corresponding error (vertical bars) are plotted

Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us