Passenger Flow Scale Prediction of Urban Rail Transit Stations Based on Multilayer Perceptron (MLP)

Table 8

Accuracy analysis of passenger flow forecast at key stations—peak hour passenger flow.

		Working days				Nonworking days
		Real value	Predictive value	AE	RE (%)	Real value	Predictive value	AE	RE (%)
Wangjing station	MLP	7582.86	6918.11	664.75	8.77	2826.00	2027.92	798.08	28.24
	RBF		6310.97	1271.88	16.77		4209.88	1383.88	48.97
	KNN		12339.29	4756.43	62.73		4593.14	1767.14	62.53
	Multiple linear regression		11698.95	4116.10	54.28		4220.07	1394.07	49.33
Xuanwu gate station	MLP	9422.2	7644.66	1777.62	18.87	2217.33	3102.19	884.85	39.91
	RBF		11541.89	2119.60	22.50		2785.65	568.32	25.63
	KNN		6880.19	2542.10	26.98		3386.19	1168.86	52.71
	Multiple linear regression		9387.40	34.89	0.37		3217.70	1000.36	45.12
Wangfujing station	MLP	6593.14	7660.82	1067.68	16.19	4467.00	5054.55	587.55	13.15
	RBF		4189.84	2403.30	36.45		2684.43	1782.57	39.91
	KNN		4278.17	2314.98	35.11		3055.43	1411.57	31.60
	Multiple linear regression		6400.83	192.31	2.92		2006.67	2460.33	55.08
Sanyuanqiao station	MLP	16947.29	9517.60	7429.69	43.84	3804.67	4069.36	264.69	6.96
	RBF		11043.97	5903.32	34.83		2809.12	995.55	26.17
	KNN		8967.41	7979.88	47.09		3244.86	559.81	14.71
	Multiple linear regression		10897.36	6049.93	35.70		3643.60	161.07	4.23
Sun palace station	MLP	7517.57	7868.74	351.17	4.67	2230.00	1960.66	269.34	12.08
	RBF		6899.90	617.67	8.22		2591.98	361.98	16.23
	KNN		7129.48	388.10	5.16		2141.86	88.14	3.95
	Multiple linear regression		7005.64	511.93	6.81		2346.11	116.11	5.21

Bold values are the values with the smallest errors in the prediction results of the four models for each site under different scenarios.