= This seminar is fully booked. If you want to participate via live stream, please send an e-mail to email@example.com =
On 16 October 2023, TRAIL PhD researcher José Ignacio Hernández will defend his PhD thesis “Data-driven Methods to Study Individual Choice Behaviour, with Applications to Discrete Choice Experiments and Participatory Value Evaluation Experiments” at Delft University of Technology. The defence is public, and you are invited to join.
Moreover, on the occasion of this public defence, TRAIL and TU Delft have set up a seminar on New Trends in Data-Driven Methods and Choice Modelling.
Two leading scholars in choice modelling will present talks on current topics in the field. In the first part, Tim Hillel presents RUMBoost, a new interpretable machine learning model inspired by random utility maximisation (RUM) models. RUMBoost combines the predictive power of Gradient-Boosted Decision Trees with the interpretability of RUM models. In the second part, Stephane Hess will address recent advancements in the estimation of advanced choice models, with a focus on an approach of avoiding poor local optima on them.
Chairman: José Ignacio Hernández
15.30 – 15.35 Opening and welcome
15.35 – 16.00 Dr. Tim Hillel: RUMBoost: Gradient boosted random utility maximisation models
16.10 – 16.35 Prof. dr. Stephane Hess: Get me out of this hole: a profile likelihood approach to avoiding poor local optima
16.45 – 17.00 Coffee and bites
TU Delft, Faculty TPM
Participation is free, registration is required: click here.
Public defence José Ignacio Hernández
At 12:30 h the public defence starts in the Senaatszaal of the Aula.
At 12:00 h José Ignacio Hernández will give a brief presentation about his research in the same room.