Discovery of Urinary Proteomic Signature for Differential Diagnosis of Acute Appendicitis

Figure 4

Construction and evaluation of diagnosis models for AA and CON. (a) ROC analysis of one-feature-based models; (b) classification model construction by 3 machine-learning algorithms: random forest (RF), support vector machine (SVM), and Naïve Bayes (NB) in the discovery dataset, and their performance evaluation by leave-one-out cross validation and the independent test.

(a)

(b)