A Hybrid Deep Learning-Based Network for Photovoltaic Power Forecasting

Table 4

Comparative analysis of the proposed model with different existing deep learning models. Herein, DKASC-AS-1A, DKASC-AS-1B, DKASC-AS-2Eco, and DKASC-Yulara-SITE-3A represent the PV power datasets. The bold text shows the experimental result of the proposed models on one-hour ahead PV power forecasting.

Dataset	Model	RMSE	MSE	MAE	MBE
DKASC-AS-1A [67]	Decision Tree	0.4531	0.2053	0.2484	0.0684
	SVR	0.4309	0.1857	0.2373	0.0463
	LSTM	0.3118	0.0972	0.1578	−0.0283
	GRU	0.3004	0.0902	0.144	0.0322
	CNN-LSTM	0.2873	0.0825	0.117	−0.0054
	CNN-GRU	0.2606	0.0679	0.1535	0.082
	LSTM-CNN	0.2239	0.0501	0.1485	−0.1472
	GRU-CNN	0.1468	0.0216	0.0742	0.0171
DKASC-AS-1B [68]	Decision Tree	0.5344	0.2856	0.3365	−0.0824
	SVR	0.5087	0.2588	0.303	0.0709
	LSTM	0.3949	0.1559	0.2219	0.0287
	GRU	0.389	0.1514	0.2064	0.0089
	CNN-LSTM	0.2776	0.0771	0.1531	0.0172
	CNN-GRU	0.262	0.0686	0.1364	−0.0318
	LSTM-CNN	0.2496	0.0623	0.208	−0.187
	GRU-CNN	0.1727	0.0298	0.0923	0.0235
DKASC-AS-2Eco [69]	Decision Tree	0.4911	0.2412	0.1909	0.0709
	SVR	0.456	0.2079	0.2246	0.0187
	LSTM	0.3167	0.1003	0.157	−0.0158
	GRU	0.3302	0.109	0.1726	−0.0176
	CNN-LSTM	0.2959	0.0876	0.1449	−0.0143
	CNN-GRU	0.2801	0.0784	0.1467	0.0132
	LSTM-CNN	0.2274	0.0517	0.1599	−0.0155
	GRU-CNN	0.1646	0.0271	0.1157	−0.0641
DKASC-Yulara-SITE-3A [70]	Decision Tree	0.416	0.173	0.2566	0.0159
	SVR	0.4966	0.2466	0.2443	−0.0122
	LSTM	0.3627	0.1315	0.1735	0.0561
	GRU	0.3864	0.1493	0.2368	−0.0013
	CNN-LSTM	0.3056	0.0934	0.1388	−0.0153
	CNN-GRU	0.3063	0.0938	0.1506	0.0354
	LSTM-CNN	0.2465	0.0608	0.155	0.0919
	GRU-CNN	0.1715	0.0294	0.1126	0.0099