This paper develops alternatively structured trip frequency/generation models, and investigates their forecast performance. The first model presented is the simple linear model with a discussion of its theoretical shortcomings. Models that address, in a progressive fashion, the underlying shortcomings of the linear model are then presented. These models are namely the truncated normal model, the Poisson model, the negative binomial model, and an ordered logit model. The modeling unit employed in the study is the individual. The models are assessed by how closely they are able to replicate trips produced by each individual in the dataset, and by each traffic zone. This assessment of performance in prediction is conducted on an estimation dataset collected in the Toronto Region in 1986, and on an independent dataset collected in the same geographic region, 10 years later, in 1996. The results show that, notwithstanding the simplicity of the simple linear model and its lack of an explicit underlying travel behavioral theory, it predicts travel in the base and forecast years with less error compared to any of the more complex models.
Forecasting Travel Demand with Alternatively Structured Models of Trip Frequency
Transportation Planning and Technology ; 30 , 5 ; 455-475
2007-10-01
21 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Forecasting Travel Demand with Alternatively Structured Models of Trip Frequency
Online Contents | 2007
|Coordinating Trip Data Between Travel Demand and Operational Models
British Library Conference Proceedings | 2005
|Passenger Travel Demand Forecasting
NTIS | 1976
Passenger travel demand forecasting
TIBKAT | 1976
|Travel Demand Forecasting 2009
Online Contents | 2009
|