Comparison of Methods Representing Heterogeneity in Logit Models

Abstract

Representing heterogeneity or taste variations in behavioural choice analysis becomes more and more an important issue in the area of transportation mode choices. Mixed logit (MXL) models have been widely adopted for this purpose. In this paper, using the data about commuting trip mode choice in the morning peak, we compare different methods that are employed to represent heterogeneity using MXL models. These methods differ in the distribution assumptions including continuous distribution, discrete mass point and a mixture of multiple continuous distributions. We review and compare the effectiveness (goodness of fit), interpretation and computational difficulty of alternative methods of representing the distribution of unobserved heterogeneity.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 7 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.