Data underlying the publication: A Logit Mixture Model Estimating the Heterogeneous Mode Choice Preferences of Shippers Based on Aggregate Data

This research proposes a way to estimate heterogeneous mode choice preferences in the context of intermodal transport directly from aggregate data. To do so, we develop a Weighted Logit Mixture (WLM) methodology , which is compared to a benchmark consisting of a Multinomial Logit (MNL) model. In the WLM, we estimate the Lognormal distribution of the intermodal cost coefficient within the shippers' population. Then, we compare the performance of our WLM against the benchmark in terms of mode share predictions, elasticities, and correlations. Finally, we also investigate the influence of adding the Value Of Time (VOT) in the WLM on the model's performance.

本研究提出了一种可直接基于汇总数据，估算多式联运场景下异质性运输方式选择偏好的方法。为达成该目标，我们构建了加权Logit混合模型（Weighted Logit Mixture, WLM），并将其与基准模型——多项Logit模型（Multinomial Logit, MNL）进行对比。在该加权Logit混合模型中，我们估算了托运人群体内多式联运成本系数的对数正态分布。随后，我们从运输方式分担率预测、弹性分析及相关性三个维度，对比了所提WLM模型与基准模型的性能表现。最后，我们还探究了在WLM模型中加入时间价值（Value Of Time, VOT）对模型性能的影响。