Length of Hospital Stay Prediction at the Admission Stage for Cardiology Patients Using Artificial Neural Network

Table 5

Results of predischarge and preadmission models for AMI and HF patients.

		⁢Predischarge model		⁢Preadmission model
		LR	ANN	LR	ANN
Accuracy (%)	No tolerance	33.91%	34.19%~36.24%	36.33%	32.99%~35.82%
	1-day tolerance	55.36%	50.16%~52.56%	55.71%	49.77%~52.82%
	2-day tolerance	66.78%	64.12%~66.07%	67.47%	63.69%~65.72%
MAE		3.76	3.83~3.91	3.76	3.87~3.97
MRE		0.69	0.71~0.74	0.72	0.73~0.77