Application to discrete choice modelling
김진석
Quantum-like approach to random utility maximization framework: Application to discrete choice modelling Jinseok Kim Technology Management, Economics, and Policy Program Graduate School of Seoul National University Seoul National University Random utility maximization (RUM) framework is the most commonly used approach to model consumer choice and decision-making processes, with the discrete choice model in particular being widely used for choice modeling. However, while the firm economic foundation of random utility theory provides accurate modelling of human behavior in most cases, there has been continuous efforts to improve upon this framework to account for human behavior under risk and uncertainty. Many studies have therefore developed expanded models to incorporate behavioral concepts into the mathematical models of choice. This study takes on this task by examining the impact of ontic uncertainty to human decision-making process. In doing so, this study explores the benefits of applying the key concepts of quantum mechanics framework to choice modelling framework in explaining human behavior under uncertainty. Through this application, this study successfully derives an altered choice probability formulation that expands upon the existing choice probability formulation with an added interference term. It is noticeable that the first term of the derived quantum-like choice probability closely resembles that of the mixed logit model. The interference term reflects the effect of applying the quantum concepts, superposition and interference effect into the discrete choice model framework. This derived formulation is then applied to empirical data to test for performance of the quantum-like discrete choice model and to examine the new implications that the derived model possesses. Estimation of parameters is implemented through the simulated maximum likelihood estimation method. Specifically, through this new methodology, this study intends to analyze and interpret the impact of uncertainty in the decision-making process. Empirical application of the quantum-like discrete choice model revealed that the quantum-like discrete choice model displayed fairly stable estimation of the true parameter values and similar parameter estimations as the mixed logit model. However, as the interference term showed to be of insignificant magnitude, the effect of ontic uncertainty and interference effect needs further examination in further application of quantum concepts to choice modelling framework. Keywords: Choice modelling, quantum probability theory, uncertainty, superposition, interference, mixed logit Student Number: 2020-36904
•Jan 1, 2024•0 citations