A closed-form multiple discrete-count extreme value (MDCNTEV) model. (October 2022)

• Record Type: Journal Article
• Authors: Bhat, Chandra R.
• Is Part Of: Transportation research. Volume 164(2022)
• Page Start: 65
• Page End: 86
• Publication Date: 2022-10
• Subjects: Multivariate count -- MDCNTEV model -- Discrete-count -- Two-stage budgeting -- Utility theory -- Tourism trip generation and destination choice; Transportation -- Research -- Periodicals; Transportation -- Mathematical models -- Periodicals
• DOI: 10.1016/j.trb.2022.08.001
• Languages: English
• ISSNs: 0191-2615

Abstract

Highlights: A new two-stage budgeting multiple discrete-count model is proposed. Alternative processes are identified for multivariate count choices. Using reverse-Gumbel stochastic terms, a closed form is obtained. A tourism trip generation and destination model is estimated. Abstract: In this paper, we propose a new two-stage budgeting-based utility-theoretic econometric multiple discrete-count model based on the linking of a fractional split MDCEV model component with a total count model. Through the strategic specification of error distributions in the model, we derive a multiple discrete-count extreme value (MDCNTEV) model that has a closed-form probability expression and that is estimable using straightforward maximum likelihood estimation. An application of the proposed model is demonstrated in the context of individuals' multivariate count of recreational episodes to each of multiple possible tourism destination locations. The results highlight the promise of the proposed model for a variety of multivariate count consumer choice settings. The model can also serve as a base model over which random heterogeneity may be superimposed to specify more advanced models.