Areas under the ROC curve (AUC) of the six variable selection methods with all simulated variables.

100 simulations were used to plot and evaluate the predictive performance of six variable selection methods when the number of total predictors (t = 100 and 200), the number of true predictors (r = 8, 12, 16 and 20) and sample size (n = 50, 100, 200, 300 and 500) change. Left panel: the number of total predictors t = 100; right panel: the number of total predictors t = 200. Variable selection methods: stepwise, stability selection, LASSO, Bolasso, two-stage hybrid and bootstrap ranking procedures.