Additional file 2 of Exploiting HIV-1 protease and reverse transcriptase cross-resistance information for improved drug resistance prediction by means of multi-label classification

Schematic illustration of the approach. The schematic setting of our MLC approach is shown for PIs. We applied binary classification for each drug using random forests and logistic regression models. The AUC values of binary classification (whereas RTV achieved the best prediction performance, APV the worst) were used to define label order in the CC. For ECCs, ensembles of thirty chains with random order were generated. For training and testing we applied a 10-fold cross-validation scheme. (PDF 20 kb)