Gabriel Nova Sepúlveda
PhD candidate in TLO | ESS | TPM at TU Delft
Member of the CityAI Lab
I am a Chilean PhD candidate in the Transport & Logistics section in the Department of Engineering Systems and Services of the faculty TPM, in the line "Using AI to automate choice modelling".
My academic background is a Civil Engineer in Transport and a Master's in Transport Engineering, both at the University of Chile. My Master's thesis, guided by Professor Angelo Guevara, focused on studying and understanding the dynamics of attribute scrutiny in discrete choice processes. Firstly, the study focused on the collection and analysis of such data using the "click-tracker" and "eye-tracker" methods mounted on a specially designed SP survey. In the second stage, the study considered formulating and validating a random utility maximisation model that considers the sequential evaluation of attributes (RUM-DFT).
Based on this research, so far, there is a predominance of breadth-first information searches in the deliberation process, this behaviour is not total but becomes more acute as the number of attributes and alternatives shown in SP surveys increases. Thus, the results suggest that the RUM model could not adequately describe the choice process. Likewise, the model that considers dynamic utilities and incorporates the assumptions behind the Decision Field Theory model, allows the recovery of both the parameters associated with the attributes and those of the information search process (tolerance and attention weights).