Optimization of Tree-Based Machine Learning Models to Predict the Length of Hospital Stay Using Genetic Algorithm

Table 7

Absolute error indices of predictive models on the test dataset.

No.	Model	Mean (day)	Median (day)	SD (day)	IQR (day)	Min (day)	Max (day)
1	KNN	1.78	1.75	1.50	1.75	0.00	6.33
2	Lm	1.56	1.33	1.14	1.56	0.00	4.56
3	Lm_transformed	1.52	1.19	1.26	1.37	0.10	5.19
4	DT_default	2.15	2.00	1.79	1.00	0.00	7.00
5	DT_GA	1.95	1.65	1.61	2.00	0.35	6.35
6	RF_default	1.65	1.02	1.47	1.87	0.09	5.71
7	RF_GA	1.76	1.34	1.42	1.50	0.04	5.57
8	XGBoost_default	2.45	2.56	1.93	2.54	0.00	7.67
9	XGBoost_GA	1.54	1.14	1.38	1.26	0.04	5.68
10	ANN	1.61	1.20	1.20	1.75	0.03	4.71