EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2985. Random Preference Model

Invited abstract in session MA-11: Choice behavior, stream Behavioural OR.

Monday, 8:30-10:00
Room: 12 (building: 116)

Authors (first author is the speaker)

1. Moha Ghaderi
Economics and Business, Pompeu Fabra University
2. Kamel Jedidi
Columbia University
3. Milosz Kadzinski
Institute of Computing Science, Poznan University of Technology

Abstract

Modeling preferences from observed choices is a pivotal theme in mainstream economics and decision theory with expansive applications in policymaking, marketing, and transportation. The standard approach in modeling preferences is to attach utility scores to the choice options consistent with the order defined by the preferences. Nevertheless, the utility scores are merely artifacts or mediums to represent preferences, whereas preferences are the underlying construct generating choices. We aim to develop a framework faithful exactly to this view.
Our model takes preferences as the central primitive. Preferences generate choice data and are represented by analytically convenient mediums called utility functions. We argue that contrary to our approach, defining preferences in terms of utilities and choices can pose severe challenges for nonparametrically inferring the underlying choice process. To address this shortcoming, we model preferences directly by a probability distribution over the set of strict linear orderings of alternatives instead of an underlying utility. We build on the theoretical ground of the random preference literature and use the distributionally robust nonparametric approach to minimize the specification and distributional assumptions required to derive robust conclusions. We address the computational challenges and demonstrate the applicability of our model through an empirical application to a conjoint setting.

Keywords

Status: accepted

Back to the list of papers