2. Modeling and Analysis of Railway Passenger Flow Forecast During the Spring Festival

The sharp increase in railway passenger flow during the Spring Festival Travel Season has tested the organization and dispatching ability of the railway transportation system. In this paper, the advantages of least square support vector machine (LSSVM) in small sample data prediction are integrated, and the ARIMA-LSSVM hybrid model based on residual linear transfer superposition is proposed, which is verified by Xiamen Spring Festival railway passenger flow. The analysis results show that the average absolute errors of hybrid model are 0.565 × 10⁴ and 0.979 × 10⁴ person times, respectively, which are 22.50% and 12.43% higher than ARIMA model, and 28.30% and 18.35% higher than LSSVM model. This study plays a positive role in improving the railway passenger flow forecasting ability and adjusting the preparation time during the Spring Festival Travel Season.