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

Table 8

The obtained results for ISBEMI Set 2.

Date	The test data	[2]	[4]	[6]	[5] distribution-based method	[5] average-based method	[11]	[29]	The proposed method
09.09.2008	40124.6	39958.4	40550	39700	40000	39967	40547	40500	40457.3
10.09.2008	39295.0	40294.2	40550	37500	40000	39800	40126	39500	40457.3
11.09.2008	37388.1	39431.8	38650	37500	40000	39100	39603	39500	39335.3
12.09.2008	37033.9	37656.1	36750	37500	37200	37500	37602	38500	39335.3
15.09.2008	35081.4	37272.4	36750	37500	37200	37100	36955	37500	36339.1
16.09.2008	33736.4	35024.0	34850	35300	34533	34700	34881	35500	33234.8
17.09.2008	32727.6	33954.4	33900	33100	34400	33700	34037	33500	33234.8
18.09.2008	32216.4	33638.4	33900	33100	32400	32700	32266	32500	33234.8
19.09.2008	36370.2	37746.4	33900	35300	32400	32300	32221	35500	36339.1
22.09.2008	36183.6	36450.8	36750	35300	36800	37700	36711	35500	36339.1
23.09.2008	35454.2	36166.4	36750	35300	36800	36100	35225	35500	36339.1
24.09.2008	35177.1	35072.6	34850	35300	35600	35100	35634	35500	33234.8
25.09.2008	36361.8	35024.0	34850	35300	34533	34700	34881	35500	36339.1
26.09.2008	36556.6	36450.8	36750	35300	36800	37700	36711	37500	36339.1
29.09.2008	36051.3	36450.8	36750	37500	36800	35900	36711	35500	36339.1
RMSE		1155.3	1229.7	1143.6	1579.1	1483.0	1484.9	1145.9	1107.2
MAPE		0.0268	0.0299	0.0261	0.0327	0.0308	0.0298	0.0255	0.0233
DA		0.4286	0.5000	0.5714	0.4286	0.7143	0.4286	0.6429	0.5714