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

Table 6

Model training accuracy analysis—peak hour passenger flow.

	Working days				Nonworking days
	MAE	RMSE	MAPE (%)	R^2	MAE	RMSE	MAPE (%)	R^2
MLP	2108.24	2844.69	36.67	0.50	855.99	1100.52	35.01	0.41
RBF	2338.57	2878.00	38.14	0.48	874.24	1255.84	35.37	0.23
KNN	2330.95	3164.72	38.11	0.38	908.98	1155.27	36.80	0.35
Multiple linear regression	1795.44	2667.71	29.04	0.56	842.59	1200.30	31.66	0.30

Bold values are the values with the smallest errors in prediction results for each model under different scenarios. It shows that the model has the least training error.