A Spatial Influence Area Model of Station Closure Based on Base Learner

In the rail transit system, station closures change the passenger flow and interrupt normal operations. It is critical to determine the influence area to ensure the safety and efficiency of unnormal operation. In this paper, a data-driven model is established from the OD (origin-destination) level to identify the influence area of station closures. Firstly, the passenger flow and travel time distribution of OD pairs are considered as the main variable parameters of the affected OD pairs. Secondly, for each variable parameter, three kinds of anomaly detection algorithms are combined based on the base learner to develop the influence area identification model. Finally, a real-world case study based on the Beijing metro network is conducted to valid the proposed method. The result shows that the proposed model has better identification effect and higher accuracy than the single anomaly detection algorithm.