Introduction:

Machine learning methods have been hot for some years and have been applied also in the wide scope of urban and transportation planning. In terms of the decision making issue, however, researchers and policy makers commonly demand a model that is both interpretable and predictable in the sense the model warrants not only a good predictability but also generates enhanced understanding of choice preferences from the model parameters. Unfortunately the emerging trend of interpretable machine learning is not always in line with such a requirement. In this presentation, I would like to bring a talk about a new model we developed in recent to extend the interpretability while at the same time enhance the model performance in terms of predictions. More specifically, we applied the model in an open data to show how the heterogeneity in the choice preferences among people can be better explained without losing its capability in predictions.

Lecturer:

Dr.Tao FENG,Professor of Hiroshima University, Japan

Host: Assoc Prof. Wen JIANG

Date:

18th November, 2022 10:30

Event Location:

Tencent Webinar  426 1834 5195 

Tencent Livestream Channel  X-scapeLab