Estimating regret-rejoice functions allowing for attribute indifference using polynomial expressions

Regret-based models have recently been introduced in transportation research as an alternative to the utilitymaximizing discrete choice models. Besides the studies examining the predictive performance of regret-based models, relative to random utility models (MNL in particular), the development of these models has focused on the specification of the regret function. While the original model only included regret, more recent models also incorporate the concept of rejoice. In addition, the curvature of the regret-rejoice function has been put under scrutiny. In this paper, we suggest a flexible specification using polynomial expressions, which does not require any a priori assumptions about the curvature of the regret-rejoice function, and in addition incorporates the notion of preference indifference.