TY - JOUR A1 - Dahmen, Victoria A1 - Weikl, Simone A1 - Bogenberger, Klaus T1 - Interpretable Machine Learning for Mode Choice Modeling on Tracking-Based Revealed Preference Data JF - Transportation Research Record: Journal of the Transportation Research Board N2 - Mode choice modeling is imperative for predicting and understanding travel behavior. For this purpose, machine learning (ML) models have increasingly been applied to stated preference and traditional self-recorded revealed preference data with promising results, particularly for extreme gradient boosting (XGBoost) and random forest (RF) models. Because of the rise in the use of tracking-based smartphone applications for recording travel behavior, we address the important and unprecedented task of testing these ML models for mode choice modeling on such data. Furthermore, as ML approaches are still criticized for leading to results that are hard to understand, we consider it essential to provide an in-depth interpretability analysis of the best-performing model. Our results show that the XGBoost and RF models far outperform a conventional multinomial logit model, both overall and for each mode. The interpretability analysis using the Shapley additive explanations approach reveals that the XGBoost model can be explained well at the overall and mode level. In addition, we demonstrate how to analyze individual predictions. Lastly, a sensitivity analysis gives insight into the relative importance of different data sources, sample size, and user involvement. We conclude that the XGBoost model performs best, while also being explainable. Insights generated by such models can be used, for instance, to predict mode choice decisions for arbitrary origin–destination pairs to see which impacts infrastructural changes would have on the mode share. KW - travel behavior KW - sensitivity analysis KW - smartphone tracking KW - mode choice KW - revealed preference KW - interpretable machine learning Y1 - 2024 U6 - https://doi.org/10.1177/03611981241246973 SN - 0361-1981 VL - 2678 IS - 11 SP - 2075 EP - 2091 PB - SAGE Publications ER - TY - INPR A1 - Takayasu, Anna A1 - Weikl, Simone A1 - Dahmen, Victoria A1 - Bogenberger, Klaus T1 - Impact of Travel Stress and Infrastructure Quality on Cycling Mode Share: Insights From Multimodal Trajectory and Survey Data Analysis N2 - This study explores the relationship between travel stress, infrastructure quality, and cycling mode share. Utilizing multimodal trajectory data, survey responses on travel stress, and detailed road network information with geological and bicycle facility data, the study analyzes correlations between actual and perceived infrastructure quality, travel stress, and cycling mode share, as well as the impact of travel time and distance on cycling mode share. The results indicate significant correlations between perceived infrastructure quality and travel stress levels, and between travel stress levels and cycling mode share, particularly among regular bike users. However, clear correlations between actual and perceived infrastructure quality, and between perceived infrastructure quality and cycling mode share, were not observed. Additionally, direct effects of travel time and distance on cycling mode share were not evident. From these findings, three main insights emerge. Firstly, accurately estimating cycling mode share requires considering multiple parameters beyond travel time and distance. Secondly, to implement effective measures for bicycle facilities, a detailed analysis of perceived infrastructure quality, incorporating actual infrastructure design, is crucial. Lastly, travel stress, when carefully assessed with consideration of various factors beyond infrastructure design, emerges as a key determinant of travel mode choice. These insights have significant implications for future cycling research, including the development of travel mode choice models and estimating cycling demand. The study's valuable contributions advance efforts to promote cycling as a sustainable and preferred mode of transportation. Y1 - 2023 U6 - https://doi.org/10.13140/RG.2.2.30696.85766 PB - Researchgate ER - TY - JOUR A1 - Dahmen, Victoria A1 - Bogenberger, Klaus A1 - Weikl, Simone T1 - Modellierung des Verkehrsmittelwahlverhaltens mittels erklärbarem Maschinellem Lernen JF - Straßenverkehrstechnik N2 - Die Modellierung der Verkehrsmittelwahl ist für die Vorhersage und das Verständnis des Mobilitätsverhaltens unerlässlich. Hierbei wurden in den letzten Jahren mit maschinellem Lernen vielversprechende Ergebnisse erzielt, insbesondere für XGBoost- und Random-Forest-Modelle. Aufgrund der zunehmenden Verwendung von Tracking-basierten Smartphone-Apps zur Aufzeichnung des Mobilitätsverhaltens, wenden wir diese Modelle auf einen solchen Datensatz an. Zudem analysieren wir eingehend deren Interpretierbarkeit. Wir kommen zu dem Schluss, dass das XGBoost-Modell am leistungsstärksten und dennoch erklärbar ist. Die von solchen Modellen gewonnenen Erkenntnisse können beispielsweise genutzt werden, um die Verkehrsmittelwahl für beliebige Quelle-Ziel-Paare vorherzusagen. Y1 - 2025 U6 - https://doi.org/10.53184/SVT2-2025-3 SN - 0039-2219 VL - 69 IS - 2 SP - 96 EP - 102 PB - Kirschbaum ER -