The overall prediction accuracy of the random forests model when classifying the memory and control group using functional connectivity, PRS, and behavioral and clinical predictors, as estimated by the AUC, was 88.29%. The 72 predictors input in the model include 29 functional connectivity, 27 personality and life experience, 12 alcohol outcomes, and 4 PRS variables (see Materials and Methods Section of the Supplementary Materials). Additional details regarding the classification accuracy are available in Section S2.2 of the Supplementary Materials.
