Fig. 2

Screening diagnostic markers based on LASSO, random forest and SVM-RFE. A The LASSO logistic regression algorithm to screen diagnostic markers. B The top 25 significant genes recognized from random forest analysis. C Feature selection based on the random forest algorithm. The x-axis indicates the number of used variables. The y-axis is the cross-validation error of each prediction model. D The intersected genes of these three analyses were selected. LASSO Least absolute shrinkage and selection operator, SVM-RFE Support vector machines-recursive feature elimination