Skip to main content

Advertisement

Fig. 1 | Arthritis Research & Therapy

Fig. 1

From: The feasibility of developing biomarkers from peripheral blood mononuclear cell RNAseq data in children with juvenile idiopathic arthritis using machine learning approaches

Fig. 1

Performance of four JIA prediction models in training and testing cohorts using all samples. a Mean accuracy, sensitivity, and specificity for four different modeling methods (KNN, RF, cSVM, gSVM) in training as assessed by tenfold cross-validation. All had accuracies > 78%. b ROC analysis in the training cohort demonstrated gSVM and RF provided best classifications with an AUC of 0.84. c Testing accuracy, sensitivity, and specificity for four models. These are true values based on the predicted class of testing samples. RF, cSVM, and gSVM had similar performance with accuracies of approximately 79%. d RF had the highest AUC (0.94) of the four models tested

Back to article page