Fig. 2From: The feasibility of developing biomarkers from peripheral blood mononuclear cell RNAseq data in children with juvenile idiopathic arthritis using machine learning approachesPerformance of four JIA prediction models in the training and testing cohorts using only European samples. a Mean accuracy, sensitivity, and specificity for four models in training as assessed by tenfold cross-validation. Accuracies ranged from 59 to 74%, with gSVM having the highest accuracy. b Similar performance is reflected in the ROC analysis of the training cohort. gSVM again had the best performance with an AUC of 0.72. c Improved accuracy, sensitivity, and specificity for four models are noted in the testing cohort as assessed by true predictions of testing samples. KNN, cSVM, and gSVM all achieved a testing accuracy of 91%. d All models had AUCs > 0.90 in the testing cohortBack to article page