A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Modelling urban freight generation using linear regression and proportional odds logit models
https://orcid.org/0000-0003-2437-0535
Abstract Receivers and intermediary establishments typical to urban areas have unique freight behaviour and are understudied, particularly in developing countries. This paper uses establishment-based freight survey data to examine their freight generation patterns in the Chennai metropolitan area, India. Three sets of Simple Linear Regression (SLR) models, three single predictor proportional odds logit models, one multiple linear regression model, and one multiple predictor proportional odds logit model with business size (employment, area, and operational age) and indicators of establishment category as regressors are developed each for both freight production and freight attraction. Partial proportional odds logit models are developed in a few cases to overcome the limitations of proportional odds logit models. The best SLR model varied with the establishment category for freight production and attraction. The establishment area model is the best among single predictor proportional odds logit models. The multiple predictor proportional odds logit models marginally improved the fit over single predictor models. The proportional odds logit model results show that the establishment category has a greater impact on freight generation levels than business size variables. Since earlier studies rarely focused on receivers and intermediary establishments, policymakers may benefit from the developed models and study insights while estimating freight demand and developing freight policies.
Highlights Receivers and intermediary establishments’ freight generation is studied. Partial Proportional Odds (PO) logit models are developed for freight attraction. Area model is the best single predictor PO logit model for freight generation. Establishment type affects freight generation more than business size variables.
Modelling urban freight generation using linear regression and proportional odds logit models
https://orcid.org/0000-0003-2437-0535
Abstract Receivers and intermediary establishments typical to urban areas have unique freight behaviour and are understudied, particularly in developing countries. This paper uses establishment-based freight survey data to examine their freight generation patterns in the Chennai metropolitan area, India. Three sets of Simple Linear Regression (SLR) models, three single predictor proportional odds logit models, one multiple linear regression model, and one multiple predictor proportional odds logit model with business size (employment, area, and operational age) and indicators of establishment category as regressors are developed each for both freight production and freight attraction. Partial proportional odds logit models are developed in a few cases to overcome the limitations of proportional odds logit models. The best SLR model varied with the establishment category for freight production and attraction. The establishment area model is the best among single predictor proportional odds logit models. The multiple predictor proportional odds logit models marginally improved the fit over single predictor models. The proportional odds logit model results show that the establishment category has a greater impact on freight generation levels than business size variables. Since earlier studies rarely focused on receivers and intermediary establishments, policymakers may benefit from the developed models and study insights while estimating freight demand and developing freight policies.
Highlights Receivers and intermediary establishments’ freight generation is studied. Partial Proportional Odds (PO) logit models are developed for freight attraction. Area model is the best single predictor PO logit model for freight generation. Establishment type affects freight generation more than business size variables.
Modelling urban freight generation using linear regression and proportional odds logit models
https://orcid.org/0000-0003-2437-0535
Abstract Receivers and intermediary establishments typical to urban areas have unique freight behaviour and are understudied, particularly in developing countries. This paper uses establishment-based freight survey data to examine their freight generation patterns in the Chennai metropolitan area, India. Three sets of Simple Linear Regression (SLR) models, three single predictor proportional odds logit models, one multiple linear regression model, and one multiple predictor proportional odds logit model with business size (employment, area, and operational age) and indicators of establishment category as regressors are developed each for both freight production and freight attraction. Partial proportional odds logit models are developed in a few cases to overcome the limitations of proportional odds logit models. The best SLR model varied with the establishment category for freight production and attraction. The establishment area model is the best among single predictor proportional odds logit models. The multiple predictor proportional odds logit models marginally improved the fit over single predictor models. The proportional odds logit model results show that the establishment category has a greater impact on freight generation levels than business size variables. Since earlier studies rarely focused on receivers and intermediary establishments, policymakers may benefit from the developed models and study insights while estimating freight demand and developing freight policies.
Highlights Receivers and intermediary establishments’ freight generation is studied. Partial Proportional Odds (PO) logit models are developed for freight attraction. Area model is the best single predictor PO logit model for freight generation. Establishment type affects freight generation more than business size variables.
Modelling urban freight generation using linear regression and proportional odds logit models
Middela, Mounisai Siddartha (author) / Ramadurai, Gitakrishnan (author)
Transport Policy ; 148 ; 145-153
2023-12-21
9 pages
Article (Journal)
Electronic Resource
English
ARTICLES - Determinants of Residential Satisfaction: Ordered Logit vs Regression Models
| Online Contents | 1999
| British Library Online Contents | 2009
Empirical Analysis of Toll-Lane Processing Times Using Proportional Odds Augmented MARS
| Online Contents | 2010