China
Africa
Explore chapters and articles related to this topic
Discrete Outcome Models
Published in Simon Washington, Matthew Karlaftis, Fred Mannering, Panagiotis Anastasopoulos, Statistical and Econometric Methods for Transportation Data Analysis, 2020
Simon Washington, Matthew Karlaftis, Fred Mannering, Panagiotis Anastasopoulos
The derivation of the multinomial probit model is provided in numerous sources, most notably Daganzo (1979). The problem with the multinomial probit is that the outcome probabilities are not closed form and estimation of the likelihood functions requires numerical integration. The difficulties of extending the probit formulation to more than two discrete outcomes have lead researchers to consider other disturbance term distributions.
Crash classification based on manner of collision: a comparative analysis
Published in Transportation Letters, 2023
Asif Mahmud, Agnimitra Sengupta, Vikash V. Gayah
This paper aims to compare the outcomes of statistical and ML models in classifying vehicular crashes based on collision type in terms of classification performances (accuracy, specificity, and sensitivity), interpretability, and generalizability. The statistical and ML models that have been considered are: Statistical models (Discrete choice models) Multinomial logit model and variantsMultinomial probit model (MNP)Machine learning (ML) models Support vector machinesTree-based ensemble learning modelsBayes classifiers
A systematic review of the factors associated with pedestrian route choice
Published in Transport Reviews, 2022
Nandita Basu, Md. Mazharul Haque, Mark King, Md. Kamruzzaman, Oscar Oviedo-Trespalacios
A wide range of statistical data analysis methods (see Figure A-2 (c)) were applied for PRC data analysis, and some studies used more than one method. Utility theory-based models, in other words, discrete choice models, were the most common data analysis technique to study tactical level PRC decisions. In this method, the optimal choice for an alternative is captured by value, which is called the utility. The decision-maker is assumed to select the alternative in the choice set with the highest utility (Ben-Akiva & Bierlaire, 2003). A range of logistic regression models were applied to examine associations between factors on the probability of choosing a given path relative to other path options considered. Common regression models include the multinomial logit model, multinomial probit model, conditional logistic regression model, and nested logit model. Other data analysis techniques included descriptive analysis, weighted index method (Wickramasinghe & Dissanayake, 2015), and Wilcoxon signed-rank test (Buliung, Larsen, Faulkner, & Stone, 2013).
Where do you live and what do you drive: Built-environmental and spatial effects on vehicle type choice and vehicle use
Published in International Journal of Sustainable Transportation, 2021
Na Chen, Gulsah Akar, Steven I. Gordon, Song Chen
The studies discussed above mostly use multivariate techniques to examine vehicle holding without considering spatial autocorrelation in the modeled relationships as a critical estimation issue in transportation modeling research (Hong et al., 2014). Spatial autocorrelation in vehicle demand analysis usually concerns with vehicle choice results at given locations being influenced by the ones at nearby locations (Black, 2010). For instance, Adjemian et al. (2010) detect and model spatial autocorrelation in vehicle type ownership and choice for the eleven vehicle categories using the 2000 San Francisco Bay Area Travel survey. The authors develop conventional OLS regression models, spatial error models and spatial lag models at regional and household levels and compare the results which reveal the existence of spatial dependence of vehicle type choice at both levels. The estimated effect of population density on choice of station wagon turns to be insignificant after adding spatial effects. Similarly, Paleti et al. (2013) develop a multinomial probit model with a spatial lag and find out that significant spatial dependency effects do exist. After controlling for spatial dependency effects, the effects of accessibility measures suggest that households prefer to acquire new cars if they are living in areas with good access to arterials and art employment. The findings from these studies highlight the importance of controlling for spatial relationship when modeling vehicle type choice.