DRG grouping by machine learning: from expert-oriented to data-based method

Table 7 Kappa coefficients of the data-based grouping trained by simulated cases

Folds	The rule generation method	Naive Bayes	SVM	CART	Random forest	XGBoost
1	0.9996	0.7141	0.7375	0.9218	0.9609	0.9680
2	0.9994	0.7104	0.7342	0.9221	0.9598	0.9710
3	0.9994	0.7067	0.7300	0.9218	0.9675	0.9688
4	0.9994	0.7069	0.7336	0.9192	0.9673	0.9712
5	0.9995	0.7101	0.7328	0.9197	0.9662	0.9669
6	0.9996	0.7064	0.7314	0.9216	0.9635	0.9709
7	0.9996	0.7126	0.7312	0.9210	0.9608	0.9676
8	0.9995	0.7067	0.7308	0.9202	0.9608	0.9719
9	0.9996	0.7067	0.7303	0.9210	0.9594	0.9703
10	0.9994	0.7120	0.7362	0.9207	0.9643	0.9680
Average	0.9995	0.7093	0.7328	0.9209	0.9631	0.9695

ISSN: 1472-6947