Ridership prediction at station level plays a critical role in subway transportation planning. Among various existing ridership prediction methods, direct demand model has been recognized as an effective approach. However, direct demand models including geographically weighted regression (GWR) have rarely been studied for local model selection in ridership prediction. In practice, acquiring insights into subway ridership under multiple influencing factors from a local perspective is important for passenger flow management and transportation planning operations adapting to local conditions. In this study, we propose an adapted geographically weighted LASSO (Ada-GWL) framework for modelling subway ridership, which involves regression-coefficient shrinkage and local model selection. It takes subway network layout into account and adopts network-based distance metric instead of Euclidean-based distance metric, making it so-called adapted to the context of subway networks. The real-world case of Shenzhen Metro is used to elaborate our proposed model. The results show that the proposed Ada-GWL model performs the best compared with the global model (ordinary least square, GWR, GWR calibrated with network-based distance metric and geographically weighted LASSO (GWL) in terms of estimation error and goodness-of-fit. Through understanding the variation of each coefficient across space (elasticities) and variables selection of each station, it provides more realistic conclusions based on local analysis. Besides, through clustering analysis of the stations according to the regression coefficients, clusters’ functional characteristics are found to be in compliance with the policy of functional land use in Shenzhen, indicating the high interpretability of Ada-GWL model from the spatial angle. In other words, the regression coefficients of different stations can provide us the local prospective to understand the influence of factors on stations’ ridership.

中文翻译：

---

一种用于预测地铁客流量的自适应地理加权 LASSO (Ada-GWL) 模型

车站级别的乘客预测在地铁交通规划中起着至关重要的作用。在现有的各种客流量预测方法中，直接需求模型已被公认为一种有效的方法。然而，包括地理加权回归 (GWR) 在内的直接需求模型很少被研究用于乘客预测中的本地模型选择。在实践中，从当地的角度了解多重影响因素下的地铁客流量，对于因地制宜的客流管理和交通规划运营具有重要意义。在这项研究中，我们提出了一种适应的地理加权 LASSO (Ada-GWL) 框架，用于对地铁乘客进行建模，其中涉及回归系数收缩和局部模型选择。它考虑到地铁网络布局，采用基于网络的距离度量而不是基于欧几里得的距离度量，使其所谓的适应地铁网络的上下文。深圳地铁的真实案例用于阐述我们提出的模型。结果表明，所提出的 Ada-GWL 模型与全局模型（普通最小二乘法、GWR、使用基于网络的距离度量和地理加权 LASSO（GWL）校准的 GWR 在估计误差和优通过了解各个系数跨空间的变化（弹性）和各台站的变量选择，在局部分析的基础上提供更真实的结论。此外，根据回归系数对台站进行聚类分析，发现集群的功能特征符合深圳市功能用地政策，表明Ada-GWL模型从空间角度具有较高的可解释性。换句话说，不同站点的回归系数可以为我们提供本地视角来了解因素对站点客流量的影响。