Short-Term Origin-Destination Based Metro Flow Prediction with Probabilistic Model Selection Approach

Table 9

The VAPE of outflow at each station on weekends.

Station	Combined	ARMA	NN	SVR	Averaging	Adaboost	kNN
1	0.99	1.6	1.06	1.22	0.99	0.94	1.09
2	1.00	1.37	1.06	1.13	1.00	0.98	0.9
3	0.86	1.65	0.92	0.87	0.85	0.92	1.06
4	0.99	1.92	1.26	1.24	0.99	1.03	1.18
5	0.75	1.54	1.45	0.97	0.75	0.74	0.68
6	0.69	3.05	1.16	0.90	0.71	0.70	0.88
7	0.81	1.74	0.98	0.89	0.81	0.82	0.99
8	0.87	1.91	1.18	1.45	0.88	0.88	1.36
9	1.00	1.64	0.86	1.74	1.04	1.09	1.95
10	0.79	2.3	1.12	1.11	0.80	0.82	1.35
11	0.84	2.43	1.29	1.17	0.84	0.88	1.38
12	0.72	2.94	3.14	1.16	0.73	0.80	1.03
13	0.77	3.2	1.5	0.92	0.79	0.94	1.12
14	0.83	2.72	1.68	1.49	0.83	0.85	0.99
15	0.73	6.28	2.12	0.72	0.74	0.8	0.84
16	0.94	2.58	1.52	1.4	0.94	0.96	0.89
17	0.89	3.17	1.26	1.06	0.89	0.88	0.87
18	1.34	1.84	1.36	1.58	1.33	1.33	1.57
19	2.13	2.73	2.20	3.86	2.11	2.07	2.22
# of best	11	0	1	1	1	5	0
Average	0.94	2.45	1.49	1.31	0.95	0.97	1.18