Fig. 3

Identification and validation of key diagnostic miRNAs in PCa. A ROC curves for DE-LMRMs in distinguishing normal from PCa tissues. B The coefficient profile of 25 DE-LMRMs in LASSO regression. C The plot displays the cross-validation error according to the log of lambda value in the LASSO analysis. D, E Plots of generalization error and prediction accuracy versus the number of features in SVM, respectively. F Feature ranking based on the XGBoost machine learning algorithm. G Venn diagram for the diagnostic miRNAs screened by LASSO regression, SVM-RFE, and XGBoost algorithms. H, I ROC curves for the diagnostic signature in the training set and validation set, respectively. J, K ROC curves for each key diagnostic miRNA in the training and validation sets