Search

• Reviewed studies using finite mixture modeling in transportation and related domains. • Identified six subdomains in transportation with the aid of nonnegative matrix factorization. • Presented a typology of eight types of heterogeneity. • Examined how the transportation literature has treated each of the key elements of the mixture modeling framework. • Discussed several methodological avenues and raised some issues that are important to mixture modeling applications. Accounting for some types of heterogeneity has been an important pathway to improving our models in the transportation domain, specifically in travel behavior research. This study examines the finite mixture modeling (latent class modeling) framework, which has been an appealing approach to that end. Through a comprehensive and systematic review, the paper aims to provide a broader understanding of the usage landscape and also insights into detailed elements. We firstly set up the mixture modeling framework; outline an arena of various relevant research fields; and explain how it is connected to transportation analyses. Then, by using the Scopus database, we explore relevant papers to investigate macroscopic trends in usage of the methodology (yearly trends and research topics). We identify six subdomains in transportation with the aid of nonnegative matrix factorization. We examine several components of the mixture modeling framework in detail. Each subsection covers certain elements of the framework and thus illuminates the landscape of usage and related issues: eight types of heterogeneity; two modeling approaches (exploratory and confirmatory); types of problems (supervised and unsupervised learning); membership model; outcome model; selecting the number of classes; comparisons with competing models; and software and estimation. At the end, we present a few current frontiers and potential directions for future research, and offer further discussion on several issues that arise in the context of mixture models. [ABSTRACT FROM AUTHOR]

Latent class choice modeling has gained great popularity in the transportation and choice modeling communities across the years. However, discussion of principles associated with the specification of the class membership model has barely appeared in the literature. Related to this issue, this study questions whether one of the basic assumptions of latent class choice modeling, that of independence between latent segmentation and the behavior generation process, is tenable. We formulate latent class choice models where the unobserved influences on latent segmentation and behavior generation are correlated, by introducing an error structure reflecting that supposition. The proposed method is applied to two empirical settings. In the first application, the dependent variable is an ordinal variable measuring willingness to share autonomous vehicle rides with strangers. In the second application, the dependent variable is a binary indicator of whether a person has used ridehailing services for social purposes. In both applications, error correlations were statistically significant, indicating that the segmentation and behavior generation processes are jointly determined. Although goodness of fits and parameter estimates per se are similar to those of the standard latent class choice models for these particular applications, allowing an error structure leads to a subtle change in model implications. In particular, our scenario analyses, which present marginal effects, illustrate the value of the proposed model for considering jointness arising from correlated errors, in contrast to standard latent class models. Lastly, we propose several avenues for future research. • Latent class model with an error structure is presented. • Provided the concept and mathematical equations for the model and marginal effects. • Model typology proposed (known/latent, exploratory/confirmatory, uncorrelated/correlated). • Explained the usefulness and practical issues of the proposed model. • Compared standard and proposed models with two empirical applications (binary & ordinal). [ABSTRACT FROM AUTHOR]

In the aftermath of the COVID-19 pandemic, teleworking is still widely adopted even though it is no longer a mandatory work arrangement. It is important to understand the motives of those who continue to telework, as diverse motives may be associated with heterogeneous travel patterns. In this study, we apply a latent class choice model to a 2021 sample of current and expected future teleworkers, and identify five latent segments based on their dominant telework-related motives, namely travel-dominant (comprising 25% of the weighted sample), flexibility-dominant (24%), career-dominant (22%), workplace-discouraged (20%), and family-dominant (9%). Specifically, travel-dominant , workplace-discouraged , and family-dominant teleworkers may use teleworking to solve particular issues, while flexibility-dominant teleworkers may simply be drawn to some of the assorted merits of teleworking (perhaps even to its "option value"). Career-dominant teleworkers tend to treat teleworking as an occasional work arrangement. We also generate detailed segment-specific profiles to better understand each type of teleworker. The segment-specific outcome models explore factors that influence the expected post-COVID teleworking frequency. We find that factors such as gender, education, and job characteristics have heterogeneous impacts on teleworkers with different dominant motives. Interestingly, results suggest gendered family-support roles even among teleworkers with apparently similarly family-oriented motivations. Overall, this study provides useful insights based on classifying teleworkers by different telework-related motives and exploring the heterogeneous impacts of other variables on teleworking frequency. The results will lay a foundation for future studies to explore teleworking impacts on post-COVID travel behavior changes. [ABSTRACT FROM AUTHOR]

Public transportation has been regarded as one of the most sustainable forms of transport, because of which local governments have put efforts in increasing the mode share of public transportation. Recently, not only level of service but also user satisfaction has been focused on for the assessment of transport policy; however, several points have been overlooked in previous research, such as attribute-based satisfaction, the relationship between level of service and satisfaction, and user expectation. This paper deals with these issues while assuming the heterogeneity of satisfaction, which is an indicator of perception. Out-of-vehicle times, particularly headway and access time, were selected in the analysis because they can affect travel behavior of current and potential public transportation customers. In this analysis, pattern-recognition models were applied, and public transportation users were segmented in terms of perception. The results suggest several notable recommendations for public transport policies and investigations. [ABSTRACT FROM AUTHOR]

This study examines potential observed and unobserved parameter heterogeneity in modeling vehicle-miles driven (VMD). This application is both important in its own right, and a vehicle for exploring conceptual, mathematical, and empirical differences among three finite-segmentation ways of addressing such heterogeneity: deterministic segmentation and (endogenous) switching for observed heterogeneity, and latent class models for unobserved heterogeneity. Based on empirical data from about 3,000 Georgia residents, we model weekly VMD, and identify key explanatory variables as well as the different sensitivities to those variables exhibited by various population segments. The study posits that people have different sensitivities to explanatory variables by residence type (urban versus "less urban"), and supporting evidence was found in the deterministic segmentation and endogenous switching models. The switching regression model characterized those who are more likely to live in urban areas, e.g. workers, pro-environmental, and those who are less favorable to driving. In addition, the estimated error correlations in the endogenous switching model corroborated that there are unobserved factors common to the joint decisions of residential choice and VMD generation. The latent class model identified lower and higher VMD-inclined segments and had a better goodness of fit than the other two models. The results confirm that (1) there is notable heterogeneity in the population with respect to the weight given to relevant factors that influence VMD, and (2) accounting for both observed and unobserved heterogeneities is useful in modeling VMD. [ABSTRACT FROM AUTHOR]