Figure 3
From: Blood-borne miRNA profile-based diagnostic classifier for lung adenocarcinoma
Construction of weighted voting diagnostic classifier using 10-fold cross validation and 10,000-times random partitioning of the training cohort.
(a) Schematic diagram showing construction of blood-borne miRNA profile-based diagnosis classifier. (b) Results of our search for the optimum number of miRNAs for diagnosis of lung adenocarcinoma, which showed 20 miRNAs with the lowest error rate. (c) Assessment of the 20 miRNA-based weighted voting classifier for diagnosis of lung adenocarcinoma. (d) ROC curve analysis of the diagnostic classifier, which showed an AUC value of 0.991. HS, healthy subjects; AD, lung adenocarcinoma; Sen, sensitivity; Spe, specificity.
