A diagnostic model for sepsis-induced acute lung injury using a consensus machine learning approach and its therapeutic implications

Table 1 Model performance of 4 classifiers in validation set

	Elastic net	Svm	Random forest	XGBoost	Ensemble
DEGs selected by model, n	27	29	20	33	53
Sensitivity	0.800	0.917	0.692	0.813	0.714
Specificity	0.792	0.852	0.750	0.714	0.789
Positive predictive value	0.706	0.733	0.643	0.684	0.714
Negative predictive value	0.864	0.958	0.789	0.833	0.789
Correct classification rate	0.795	0.872	0.730	0.757	0.758
AUC	0.781	0.846	0.727	0.731	0.876

ISSN: 1479-5876