Table 2.

Average forecasting performance (accuracy) of the smart persistence model and the stacked LSTM models with and without TL.

PV	Metric	Without TL	Smart Persistence	TL Strategy 1	TL Strategy 2	TL Strategy 3
$P{V}_2$	RMSE (KWh)	3.77	3.27	2.34	2.37	2.36
	MBE (KWh)	$-$ 0.25	0.17	0.16	0.06	0.05
	MAE (KWh)	1.81	1.38	1.11	1.11	1.12
	nRMSE (%)	53.14	46.18	33.06	33.38	33.28
	$R^2$ (%)	87.74	91.37	95.57	95.49	95.51
$P{V}_3$	RMSE	11.39	17.32	10.59	11.35	11.48
	MBE	0.83	0.02	0.79	0.81	0.80
	MAE	6.01	6.49	5.29	5.94	5.85
	nRMSE	40.78	62.15	37.94	40.63	41.13
	$R^2$	93.60	85.27	94.49	93.68	93.53
$P{V}_4$	RMSE	1.65	2.69	1.48	1.46	1.46
	MBE	$-$ 0.01	$-$ 0.82	0.11	0.06	0.07
	MAE	0.81	1.32	0.70	0.69	0.70
	nRMSE	35.57	57.84	31.94	31.42	31.51
	$R^2$	94.46	85.40	95.55	95.70	95.68
$P{V}_5$	RMSE	2.66	5.27	2.31	2.31	2.31
	MBE	$-$ 0.08	0.35	$-$ 0.04	$-$ 0.11	$-$ 0.16
	MAE	1.37	1.52	1.13	1.15	1.15
	nRMSE	31.68	62.75	27.56	27.46	27.54
	$R^2$	95.17	81.51	96.43	96.46	96.44
$P{V}_6$	RMSE	6.75	8.59	5.65	5.50	5.51
	MBE	1.35	$-$ 0.04	0.94	0.43	0.56
	MAE	3.54	4.25	2.74	2.79	2.77
	nRMSE	40.92	52.11	34.23	33.37	33.40
	$R^2$	92.72	88.58	95.07	95.33	95.31
$P{V}_7$	RMSE	3.64	4.59	2.47	2.40	2.41
	MBE	$-$ 0.16	0.15	0.24	$-$ 0.02	0.03
	MAE	1.75	1.64	1.15	1.11	1.12
	nRMSE	40.87	51.49	27.78	26.97	27.08
	$R^2$	92.23	88.19	96.56	96.76	96.74

RMSE, MBE and MAE are measured in KWh, while for $R^2$ and nRMSE the percentage is given for each model