Machine Learning and Travel Choice

Next Monday, March 4 at 16:30 p.m., José Ángel Martín Baos, professor of ESI at UCLM, will give an online seminar within the program “Machine Learning NeEDS Mathematical Optimization"(https://congreso.us.es/mlneedsmo/) organized by Emilio Carrizosa (IMUS – Institute of Mathematics of the University of Seville) and Dolores Romero Morales (CBS – Copenhagen Business School). The conference will be broadcast online through the link https://eu.bbcollab.com/guest/f36b823fbfc74849848d66808d8db459

The title of the conference is “Can machine learning methods effectively model travel mode choice? Beyond predictive performance”. It will address how machine learning models can be used to predict the choice of travel mode, highlighting the importance of factors beyond predictive performance, such as interpretability, computational complexity and efficiency. Limitations of previous research are addressed through a systematic comparison of various models in multiple aspects, showing that the models with the best predictive performance in the literature (such as Gradient Boosting and Random Forests) often offer worse estimates of behavioral indicators and aggregate market shares. , compared to alternatives such as Deep Neural Networks and Multinomial Logit (MNL) models.