Development, comparison, and internal validation of prediction models to determine the visual prognosis of patients with open globe injuries using machine learning approaches

Table 6 Predictive performance of classifiers for the final VA of OGI patients

Models Name	AUC-ROC	AUC-PRC	PPV	Sensitivity	Accuracy	F1	MCC	BS
SVM	0.86	0.75	0.812	0.764	0.885	0.787	0.709	0.433
Naïve Bayas	0.84	0.72	0.857	0.705	0.885	0.774	0.704	0.634
XGB	0.88	0.82	0.923	0.705	0.901	0.800	0.744	0.395
ADA	0.86	0.74	0.750	0.705	0.852	0.727	0.626	0.643
Bagging	0.89	0.79	0.764	0.764	0.868	0.764	0.673	0.412
Multinomial LR	0.84	0.75	0.866	0.764	0.901	0.812	0.748	0.439
KNN	0.81	0.69	0.750	0.705	0.852	0.727	0.626	0.476
Decision Tree	0.85	0.75	0.785	0.647	0.852	0.709	0.617	0.413
RF	0.92	0.86	0.866	0.764	0.901	0.812	0.748	0.311
ANN	0.96	0.91	0.894	0.819	0.930	0.855	0.811	0.201

* Area Under Curve of Receiver Operator Characteristic (AUC-ROC), Area Under Curve of Precision-Recall (AUC-PRC), Precision (PPV), Sensitivity (Recall), Accuracy (Acc), F-measure (F1), Matthew’s Correlation Coefficient (MCC), Brier Score (BS).

ISSN: 1472-6947