2019 年 13 巻 p. 503-522
The combination of multi-mode on a trip appears significantly in the total amount of trip generated daily. Understanding the influential mechanism in the multi-mode choice process will greatly help policymaking. This work introduces the application of Gradient Boosting Machine, a Machine Learning algorithm, and Local Interpretable Model-agnostic Explanations technique to investigate the multi-mode pattern and its determinants in order to deeper understand this problem. The empirical results from the case study of Jakarta city show that the single-mode choice was affected by limit features while the multi-mode choice was influenced by the wide range of variables. Additionally, the Gradient Boosting Machine was indicated with the impressive potentiality in solving this subject not only by its performance but the effectiveness in dealing with the big data.
