基于随机森林回归算法的高速铁路短期客流预测研究

Study on Forecast of High-speed Railway Short-term Passenger Flow based on Random Forest Regression

客流短期预测是高速铁路客运运营的重要依据。在阐述高速铁路短期客流预测常用方法、分析短期客流影响因素及数据处理的基础上,基于随机森林回归算法构建短期客流预测模型,并结合OOB残差均方对客流影响因素的重要性进行评估。以京沪高速铁路为例,采用2015年7月至8月北京至上海的客流量数据进行验证。结果表明,每日客流的预测精度为0.92,出发日期和运行时间是影响短期客流的重要因素。

The forecast of high-speed railway short-term passenger flow is an important reference for high-speed railway passenger transport operation. On the basis of expounding the common forecast methods of high-speed railway short-term passenger flow, and analyzing the main factors influencing short-term passenger flow and its data processing, the forecast model of short-term passenger flow is established based on random forest regression, and combining with OOB residual mean square, the importance of the factors influencing the passenger flow is evaluated. Taking Beijing-Shanghai high-speed railway as an example, the forecast model is validated through using Beijing-Shanghai passenger flow data from July to August in 2015. The result shows that the forecast accuracy of daily passenger flow is 0.92, and the departure date and running time are the most important factors influencing the short-term passenger flow.