Skip to main content

Advertisement

Table 2 Performance of baseline models

From: Machine learning to predict rapid progression of carotid atherosclerosis in patients with impaired glucose tolerance

Performance parameter Ml Naïve Bayes with feature selection Ml Naïve Bayes without feature selection Multilayer perceptron with feature selection Multilayer perceptron without feature selection Random forest with feature selection Random forest without feature selection
AUC 0.797 0.745 0.711 0.703 0.736 0.703
Correctly classified cases 340 (89.2 %) 290 (75.9 %) 339 (88.7 %) 330 (86.4 %) 338 (88.5 %) 343 (89.8 %)
Incorrectly classified cases 42 (10.8 %) 92 (24.1 %) 43 (11.3 %) 52 (13.6 %) 44 (11.5 %) 39 (10.2 %)
Brier score 0.085 0.222 0.086   0.105