An ARMA Type Fuzzy Time Series Forecasting Method Based on Particle Swarm Optimization

Table 10

The obtained results (last 15 observations are test set) for ISBEMI Set 3.

Date	The test data	[2]	[4]	[5] distribution-based method	[5] average-based method	[11]	[32]	The proposed method
11.12.2009	49386	49872	50250	49500	49100	49748	49516	49982.39
14.12.2009	50198	48606	48750	49500	49300	49316	50064	49982.39
15.12.2009	50450	49872	50250	49900	50500	50405	50942	50833.82
16.12.2009	50817	50294	50250	49900	48900	48886	50217	50833.82
17.12.2009	49963	50294	50250	50300	50900	48886	49641	50833.82
18.12.2009	50138	49872	50250	49900	49900	49748	49619	49613.83
21.12.2009	51281	49872	50250	49900	50500	50405	50932	51390.68
22.12.2009	51533	51137	51000	50300	50967	50625	50817	51390.68
23.12.2009	51162	51137	51000	51900	51500	51065	51100	51390.68
24.12.2009	51461	51137	51000	50300	50550	50625	50646	51390.68
25.12.2009	51661	51137	51000	50300	51500	51065	51073	51390.68
28.12.2009	51619	51137	51000	51900	51700	51065	51117	51390.68
29.12.2009	51786	51137	51000	51900	51700	51065	51114	51390.68
30.12.2009	51668	51137	51000	51900	51700	51963	51119	51390.68
31.12.2009	52825	51137	51000	51900	51700	51065	51117	51390.68
RMSE		810.8493	820.5748	815.9879	760.7710	917.1070	662.3497	522.0852
MAPE		0.0128	0.0133	0.0134	0.0110	0.0147	0.0106	0.0075
DA		0.2857	0.5000	0.2857	0.5000	0.3571	0.5000	0.6429