Functional Data Approach for Short-Term Electricity Demand Forecasting

Table 2

Monthly forecasting errors for the electricity demand using FAR (p), FARX (p), and AR (7) models.

Month	Errors	FAR (p)	FARX (p)	AR (7)
January	MAE	1976.277	1706.831	2143.758
	MAPE	4.013	3.514	7.303
	RMSE	2571.641	2382.379	2873.344
February	MAE	1654.44	1200.22	2055.76
	MAPE	3.497	2.523	5.461
	RMSE	1940.057	1474.227	2368.534
March	MAE	1455.708	1269.283	1717.423
	MAPE	3.276	2.804	5.292
	RMSE	1843.277	1694.025	2818.67
April	MAE	1675.477	1352.491	2081.115
	MAPE	5.338	3.435	6.442
	RMSE	1802.199	1223.685	2812.542
May	MAE	1485.393	1099.251	1715.463
	MAPE	7.505	2.992	9.153
	RMSE	1717.73	1513.166	2055.535
June	MAE	1492.786	996.807	1957.983
	MAPE	8.069	2.764	11.124
	RMSE	1462.905	1263.867	1795.036
July	MAE	1359.679	817.073	2292.06
	MAPE	8.208	2.343	10.539
	RMSE	1496.467	1074.037	1853.465
August	MAE	1430.913	752.507	1876.894
	MAPE	7.646	2.189	11.899
	RMSE	1322.593	959.575	2393.628
September	MAE	1305.892	821.495	1789.749
	MAPE	6.912	2.481	10.713
	RMSE	1584.888	1066.227	2709.673
October	MAE	1332.144	950.921	1833.398
	MAPE	5.402	2.341	7.256
	RMSE	1991.167	1215.694	2765.328
November	MAE	1374.963	1017.936	1874.082
	MAPE	3.109	2.241	5.673
	RMSE	1674.714	1291.18	1874.082
December	MAE	1848.797	1569.336	2081.957
	MAPE	5.051	3.405	9.978
	RMSE	2314.847	2133.82	2987.564