highly significantP = 9.99 × 10 −5
Group classification using the support vector machine (SVM) Overall, the set of 12 variables achieved good classification accuracy (81%), sensitivity (78%), and specificity (85%), and each of these were highly significant compared to chance simulations using permutation tests (accuracy P = 9.99 × 10 −5 ; sensitivity P = 9.99 × 10 −5 ; specificity P = 6.99 × 10 −4 ).