Machine Learning-Based Multiomics Prediction Model for Radiation Pneumonitis

Table 2

The evaluation indicators of four ML models under 11 classifiers (mean ± standard deviation).

Models	Training AUC	Testing AUC	Accuracy	Precision	Recall	F1
DVH	0.804 ± 0.053	0.801 ± 0.06	0.751 ± 0.068	0.738 ± 0.067	0.783 ± 0.16	0.751 ± 0.095
Radio + DVH	0.856 ± 0.044	0.805 ± 0.046	0.809 ± 0.062	0.772 ± 0.066	0.886 ± 0.118	0.821 ± 0.067
Radio + dose	0.833 ± 0.041	0.801 ± 0.057	0.806 ± 0.048	0.756 ± 0.045	0.911 ± 0.128	0.821 ± 0.061
Radio + eqdose	0.900 ± 0.033	0.799 ± 0.050	0.856 ± 0.034	0.809 ± 0.070	0.950 ± 0.059	0.870 ± 0.024