Figure 2.

ROC curves based on multivariable stepwise penalized logistic regression models (stepPLR) adjusting for gender effect. Models that are also adjusted for age effect outperformed those that did not control for age, across all choices of the parameter λ. The age-adjusted final model for λ = 0.1 showed the best performance in terms of AUC. See Table 1 for detailed results and the text for discussion of these results.

Wang et al. BMC Medical Genomics 2013 6(Suppl 3):S9   doi:10.1186/1755-8794-6-S3-S9