Short-Term Passenger Flow Forecast of Rail Transit Station Based on MIC Feature Selection and ST-LightGBM considering Transfer Passenger Flow

Table 4

The MAE () and MSE () results of four experiments (listed mean and standard deviation are averaged over 5 folds).

Experiment	Method
1	SARIMA	8.22 ± 0.51	161.16 ± 8.69
2	SVR (without feature selection)	9.65 ± 0.68	173.36 ± 9.62
	SVR (with feature selection)	8.52 ± 0.54	152.54 ± 7.92
3	BP network (without feature selection)	9.42 ± 0.49	181.35 ± 8.46
	BP network (with feature selection)	8.36 ± 0.42	154.22 ± 6.51
4	ST-LightGBM (without feature selection)	6.93 ± 0.13	118.27 ± 3.42
	ST-LightGBM (with feature selection)	5.72 ± 0.11	85.76 ± 2.42