Table 3.

Prediction performance of weekly emergency department patient arrivals.

Models		Training			Testing
		MAPEa (%)	RMSEb	MAPE (%)		RMSE
Patient arrivals
	ARIMAc	3.6	17.02	5.2		24.28
	ANNd	3.5	19.19	3.6		19.20
	SVMe	2.2	18.52	4.1		18.59
	RFf	2.5	19.36	3.0		19.67
	LSTMg	2.9	16.93	4.2		20.79
	ELMh	2.8	16.52	3.2		16.99
Patient arrivals + traditional variables
	GLMi	3.2	16.79	5.3		24.44
	ARIMAXj	3.5	17.85	5.1		23.16
	ANN	3.4	16.10	4.0		18.48
	SVM	2.8	16.24	3.9		17.45
	RF	2.9	17.05	3.7		18.36
	LSTM	3.2	16.43	4.4		19.94
	ELM	2.7	13.17	3.5		16.72
Patient arrivals + traditional variables + internet search index
	GLM	3.2	16.24	5.0		23.18
	ARIMAX	3.4	17.84	5.1		22.00
	ANN	3.0	14.51	3.3		15.45
	SVM	2.6	14.84	3.1		15.09
	RF	2.9	15.92	3.3		16.32
	LSTM	3.0	15.15	3.4		16.69
	ELM	2.6	13.10	3.0		14.55

^aMAPE: average mean absolute percentage error.

^bRMSE: root mean square error.

^cARIMA: autoregressive integrated moving average model.

^dANN: artificial neural network.

^eSVM: support vector machine.

^fRF: random forest.

^gLSTM: long short-term memory.

^hELM: extreme learning machine.

ⁱGLM: generalized linear model.

^jARIMAX: ARIMA with explanatory variables.