Predicting the incidence of infectious diarrhea with symptom surveillance data using a stacking-based ensembled model

Table 2 The minimum RMSEs, MAEs and MAPEs of testing set for ANN, LSTM, SVR, and XGBoost and their included variables

Model	Variables included in the model	RMSE	MAE	MAPE
ANN	N_cases	95.52	66.56	18.20%
	N_cases, N_{clinic-visits} and Rate_clinic	86.57	60.47	16.83%
	N_cases, T, T_min, T_max, Pressure, RH, and Precipitation	90.65	69.91	19.46%
	N_cases, T, T_min,T_max, Pressure, RH, Precipitation, N_{clinic-visits} and Rate_clinic	107.03	78.99	20.42%
LSTM	N_cases	88.67	57.50	16.38%
	N_cases, N_{clinic-visits} and Rate_clinic	84.85	59.07	15.92%
	N_cases, T, T_min, T_max, Pressure, RH, and Precipitation	92.71	68.21	20.13%
	N_cases, T, T_min, T_max, Pressure, RH, Precipitation, N_{clinic-visits} and Rate_clinic	95.35	66.25	18.37%
SVR	N_cases	91.78	61.74	16.55%
	N_cases, N_{clinic-visits} and Rate_clinic	86.41	61.31	16.37%
	N_cases, T, T_min, T_max, Pressure, RH, and Precipitation	121.04	92.26	29.22%
	N_cases, T, T_min, T_max, Pressure, RH, Precipitation, N_{clinic-visits} and Rate_clinic	121.38	93.70	29.42%
XGBoost	N_cases	93.43	63.12	17.31%
	N_cases, N_{clinic-visits} and Rate_clinic	89.19	66.98	18.70%
	N_cases, T, T_min, T_max, Pressure, RH, and Precipitation	88.00	63.61	17.25%
	N_cases, T, T_min, T_max, Pressure, RH, Precipitation, N_{clinic-visits} and Rate_clinic	85.13	63.77	18.76%

ISSN: 1471-2334