Figure 2
From: Computational identification of multi-omic correlates of anticancer therapeutic response
Mean prediction performance of multi-omic drug response signatures generated using elastic net regression, a support vector machine, or random forest on the Cancer Genome Project (CGP) data. Prediction performances (precision) are quantified as the proportion of true positive drug sensitive classifications to all positive classifications. Error bars represent the standard deviation of the precisions calculated during ten repetitions of ten-fold cross validation.