The Combination Forecasting of Electricity Price Based on Price Spikes Processing: A Case Study in South Australia

Table 2

The forecasting results of three individual models.

Time	Actual price value ($/MWh)	ARIMA		BP		Kalman
		Forecasting value ($/MWh)	MAPE (%)	Forecasting value ($/MWh)	MAPE (%)	Forecasting value ($/MWh)	MAPE (%)
00:30	41.91	22.04	47.41	17.77	57.61	24.58	41.34
01:00	37.92	19.25	49.24	18.43	51.40	20.04	47.15
01:30	34.93	20.99	39.90	20.61	40.99	23.86	31.70
02:00	24.75	18.54	25.07	18.45	25.45	21.34	13.79
02:30	22.86	15.42	32.53	10.50	54.07	16.54	27.66
03:00	20.49	15.33	25.17	17.59	14.16	17.24	15.87
03:30	18.55	15.33	17.35	17.59	5.18	17.24	7.07
04:00	18.82	15.82	15.92	17.82	5.31	16.95	9.92
04:30	19.59	18.44	5.86	18.32	6.46	19.09	2.57
05:00	19.57	20.31	3.80	19.27	1.52	20.55	5.00
05:30	21.61	22.51	4.15	19.77	8.50	22.70	5.03
06:00	21.14	24.63	16.50	26.05	23.24	24.41	15.47
06:30	21.63	20.51	5.16	34.76	60.72	30.35	40.33
07:00	21.63	47.05	117.54	26.16	20.94	36.19	67.31
07:30	19.85	25.52	28.54	26.72	34.60	28.81	45.12
08:00	21.66	27.48	26.88	32.36	49.38	30.00	38.50
08:30	30.75	30.15	1.94	34.72	12.91	32.48	5.61
09:00	31.95	29.32	8.23	30.23	5.39	38.02	19.00
09:30	32.57	27.89	14.37	28.75	11.74	37.98	16.62
10:00	40.81	28.76	29.52	28.44	30.31	40.39	1.03
10:30	44.21	28.91	34.61	32.05	27.51	38.72	12.42
11:00	43.97	26.24	40.32	32.49	26.10	39.26	10.71
11:30	44.66	31.46	29.55	35.30	20.97	36.95	17.26
12:00	43.25	40.80	5.66	37.51	13.27	38.74	10.42