A nested recursive logit model for route choice analysis

Mai, Tien; Fosgerau, Mogens; Frejinger, Emma

doi:10.1016/j.trb.2015.03.015

Cited by 114 publications

(80 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The only novelty that we introduce is that we specify scale parameters in one of the NL models as an exponential function. The same trick is used by Mai et al (2015) to allow scale parameters to vary while avoiding the estimation of a large number of parameters.…”

Section: Choice Modelsmentioning

confidence: 99%

An empirical study on aggregation of alternatives and its influence on prediction in car type choice models

2017

Self Cite

View full text Add to dashboard Cite

Assessing and predicting car type choices are important for policy analysis. Car type choice models are often based on aggregate alternatives. This is due to the fact that analysts typically do not observe choices at the detailed level that they are made. In this paper, we use registry data of all new car purchases in Sweden for two years where cars are observed by their brand, model and fuel type. However, the choices are made at a more detailed level. Hence, an aggregate (observed) alternative can correspond to several disaggregate (detailed) alternatives. We present an extensive empirical study analyzing estimation results, in-sample and out-of-sample fit as well as prediction performance of five model specifications. These models use different aggregation methods from the literature. We propose a specification of a two-level nested logit model that captures correlation between aggregate and disaggregate alternatives. The nest specific scale parameters are defined as parameterized exponential functions to keep the number of parameters reasonable. The results show that the in-sample and out-of-sample fit as well as the prediction performance differ. The best model accounts for the heterogeneity over disaggregate alternatives as well as the correlation between both disaggregate and aggregate alternatives. It outperforms the commonly used aggregation method of simply including a size measure.

show abstract

Section: Choice Modelsmentioning

confidence: 99%

An empirical study on aggregation of alternatives and its influence on prediction in car type choice models

2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…Various extensions of the RL have been proposed (Mai et al, 2015;Mai, 2016;Mai et al, 2016). Currently, we are working on extending and generalizing the MRI model.…”

Section: Resultsmentioning

confidence: 99%

“…Both the RL and the MRI models can be extended to account for the correlation among the alternatives. We refer the reader to Mai et al (2015), Mai (2016) and Mai et al (2016) for the extensions of the RL model accounting for correlation using nested, cross-nested and mixed logit formulations, respectively. In the following paragraphs, we describe how the MRI model addresses the correlation of alternatives.…”

Section: Addressing the Correlation Of Alternativesmentioning

confidence: 99%

“…The proposed technique avoids the full enumeration of paths and does not require sampling of alternatives. It is further extended by Mai et al (2015), Mai (2016) and Mai et al (2016) to account for correlation of alternatives by means of a nested, cross-nested and mixed logit formulation, respectively.…”

Section: Introductionmentioning

confidence: 99%

“…Both models can be extended to account for correlation among the alternatives. This has been done for the RL model (Mai et al, 2015;Mai, 2016 andMai et al, 2016). We describe how this generalization applies to the MRI model.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Operational route choice methodologies for practical applications

2018

View full text Add to dashboard Cite

Understanding the route choice of individuals is a key question in travel behavior research and fundamental to transportation demand management. Modeling route choice involves several challenges, pertaining to both data availability and accuracy, and model structure and computational efficiency. The main challenge related to the route choice problem, namely the generation of the choice set, has been circumvented by Fosgerau et al. (2013) through a Markovian link choice problem that is equivalent to a path-based logit model. It is called the recursive logit (RL) model and it is directly applicable for analysis at a disaggregate level. Recent methodologies are designed to simplify the structure of the model and reduce the data requirements. Kazagli et al. (2016) propose a representation of routes that is based on aggregate elements, inspired by the environmental images of the physical world formed by the individuals (Lynch, 1960), instead of links. It is called the mental representation items (MRI) model and is useful for analysis at an aggregate level. From an application point of view, both disaggregate and aggregate route choice indicators are useful. In this paper, we describe how the RL and the MRI models can be used in practice to derive link and route flows, that are the most relevant indicators for route choice applications. Our analysis identifies the advantages and the limitations of each model and allows to draw insights into the use of a specific model, depending on the needs of the application and the data availability. Finally, we investigate the performance of the two models on real data and discuss specific features of the MRI model. The results demonstrate the validity of the MRI model for the purposes of an aggregate analysis.

show abstract