Development and validation of three machine-learning models for predicting multiple organ failure in moderately severe and severe acute pancreatitis

Table 2 Different combinations of features by SVM

Combination of features No. of features	HCT	PT	APTT	TT	FIB	D-dimer	R-time	K-time	α	MA	CI	CRP	IL-6	PCT	BUN	Creatinine	AUC
1															√		0.7015
2			√												√		0.7690
3			√										√		√		0.8006
4	√										√		√			√	0.8130
5	√						√	√					√			√	0.8168
6	√				√		√					√	√			√	0.8278
7	√				√					√	√	√	√			√	0.8362
8	√				√		√			√	√	√	√			√	0.8378
9	√				√	√	√	√			√	√	√			√	0.8396
10	√				√	√	√	√			√	√	√		√	√	0.8370
11	√				√	√	√	√		√	√	√	√	√		√	0.8382
12	√		√	√	√	√	√	√		√		√	√		√	√	0.8338
13	√	√	√	√	√	√	√	√		√		√	√		√	√	0.8332
14	√		√	√	√	√	√	√		√	√	√	√	√	√	√	0.8301
15	√	√	√	√	√	√	√	√		√	√	√	√	√	√	√	0.8250
16	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	√	0.8157

ISSN: 1471-230X