Hybrid Time-Series Prediction Method Based on Entropy Fusion Feature

Table 5

Comparative experimental results of different models.

Index	Model	Power consumption (kW h)	Temperature (°C)	Wind speed (M/s)	Humidity (%)
RMSE	EEMD-ELSTM	575.15	0.51	0.0038	1.29
	LSTM	4739.34	4.34	0.0068	7.85
	ARIMA	1217.73	0.72	0.0054	4.98
	SVR	7206.11	3.42	0.0344	8.91
	XGBoost	866.60	0.55	0.0051	1.91
	Prophet	10793.95	6.67	0.0079	15.25
MAE	EEMD-ELSTM	449.06	0.35	0.0026	0.99
	LSTM	4366.71	4.43	0.0042	7.38
	ARIMA	928.45	0.49	0.0041	2.62
	SVR	6270.58	3.32	0.0339	8.83
	XGBoost	666.23	0.38	0.0032	1.43
	Prophet	9777.84	6.08	0.0065	11.85
MAPE (%)	EEMD-ELSTM	1.29	2.47	3.04	1.27
	LSTM	12.44	27.41	4.73	9.08
	ARIMA	2.42	4.92	5.04	3.89
	SVR	15.15	18.38	43.48	12.06
	XGBoost	2.10	2.48	3.623	1.79
	Prophet	33.31	48.03	7.68	13.71