AUC Values of ROC Curve Analysis Based on the Selected MPAs Variables to Discriminate Different Subgroups in the Training and Validation Sets Using Different Machine Learning Algorithms: (A) Random Forest Models Based on the Training Sets, (B) Random Forest Models Based on the Validation Sets, (C) SVM Models Based on the Training Sets, (D) SVM Models Based on the Validation Sets, (E) XGBoost Models Based on the Training Sets, (F) XGBoost Models Based on the Validation Set. Abbreviations: AOS, adult-onset schizophrenia; EOS, early-onset schizophrenia; HC, healthy controls; ROC curve, receiver operating characteristic curve; SVM, support vector machine; XGBoost, eXtreme gradient boosting.