Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: Predicting combinative drug pairs towards realistic screening via integrating heterogeneous features

Fig. 2

The flowchart of predicting drug combination by integrating heterogeneous sources of drugs. The pair of d x and d y is input into three classifier models, which are trained by three kinds of features of drug pairs, including DDI, DTI and SE. The confidence scores of the pair being a potential drug pair are reported by those classifiers and are further integrated with its ATC-based similarity entry. The average of these scores is taken as the final confidence score of indicating how likely the pair can be a combinative drug pair

Back to article page