Fig. 3From: Integrating gene set analysis and nonlinear predictive modeling of disease phenotypes using a Bayesian multitask formulationAUROC values on the cancer datasets for MSI-H versus others classification. The box-and-whisker plot shows the results over 100 replications in repeated random subsampling validation experiments of BRVM, SBMKL and SBMTMKL on both datasets. The numbers just below the dataset names give the average AUROC value for each experimentBack to article page