Machine Learning Readmission Risk Modeling: A Pediatric Case Study

Table 6

Average standard deviation AUC performance of SVM, MLP1, MLP2, and NB for decreasing number of folders in the RCV process. no SMOTE = no oversampling correction of class imbalance is done.

nfolds	SMOTE
	SVM	MLP2	MLP1	NB
10	0.597 0.022	0.539 0.022	0.643 0.020	0.654 0.014
5	0.587 0.010	0.55 0.018	0.634 0.011	0.653 0.014
4	0.585 0.008	0.548 0.021	0.63 0.009	0.655 0.008
3	0.584 0.009	0.55 0.011	0.628 0.010	0.653 0.011
	no SMOTE
	SVM	MLP2	MLP1	NB
10	0.495 0.020	0.631 0.026	0.661 0.021	0.656 0.014
5	0.481 0.019	0.615 0.008	0.661 0.008	0.658 0.007
4	0.473 0.004	0.631 0.011	0.661 0.012	0.659 0.008
3	0.471 0.007	0.627 0.015	0.657 0.002	0.658 0.009