A stacking-based model for predicting 30-day all-cause hospital readmissions of patients with acute myocardial infarction

Table 5 Comparison of our study and previous works

Author	Balance method	Feature selection	Samples	Variables	IR	Method	AUC
Krumholz [52]	no	Stepwise Logistic Regression	200,750	103	1:4.29	LR	0.630
Yu [53]	no	no	844	unknown	1:3.76	linear-SVM	0.660
Gupta [15]	no	no	7018	192	1:5.12	GBDT	0.641
Ours	NCR	SFM	3283	293	1:6.72	Stacking	0.720

IR Imbalance ratio, NCR Neighborhood clean rule, SFM SelectFromModel, LR Logistic regression, linear-SVM, Linear support vector machine, GBDT Gradient boosting decision tree

ISSN: 1472-6947