Short-Term Power Load Forecasting of Integrated Energy System Based on Attention-CNN-DBILSTM

Table 2

Comparison of prediction accuracy evaluation indicators.

Date	Evaluation index	Result
		CNN-LSTM	CNN-BILSTM	CNN-DBILSTM	Attention-CNN-DBILSTM	Attention-CNN-DBILSTM considering multiple load correlation
5.7–5.13	MAPE (%)	6.28	6.07	5.08	5.37	4.36
	RMSE (kW)	718.45	682.24	669.59	577.23	538.67
6.6–6.12	MAPE (%)	6.78	6.12	5.89	5.34	3.77
	RMSE (kW)	732.12	673.32	591.21	563.20	522.69
7.13–7.19	MAPE (%)	6.32	5.89	5.77	5.43	3.98
	RMSE (kW)	693.87	632.51	588.82	569.32	502.07
8.3–8.9	MAPE (%)	6.17	6.03	5.76	4.78	3.94
	RMSE (kW)	701.09	656.01	621.22	587.91	531.73
9.12–9.18	MAPE (%)	6.12	5.67	5.54	4.50	4.11
	RMSE (kW)	698.34	636.99	598.18	543.43	523.97
10.13–10.20	MAPE (%)	6.84	6.49	5.92	4.97	3.48
	RMSE (kW)	722.45	680.65	617.10	598.32	541.54
11.21–11.27	MAPE (%)	6.88	6.33	5.65	4.91	4.21
	RMSE (kW)	721.31	670.82	621.15	585.19	539.86
12.2–12.8	MAPE (%)	6.87	5.91	5.26	5.13	3.51
	RMSE (kW)	691.03	673.64	622.88	580.96	562.37