MODELLING PERCEIVED QUALITY FOR URBAN TRANSPORT SYSTEMS USING WEIGHTED VARIABLES AND RANDOM PARAMETERS
[EN] In this article, an Ordered Logit model is proposed considering systematic and random variations in tastes. The methodology followed for the creation of this model consisted, in first place, in obtaining data using a revelled preferences survey. In the survey, each user had to evaluate, followi...
| Autores: | , , |
|---|---|
| Tipo de recurso: | capítulo de libro |
| Fecha de publicación: | 2016 |
| País: | España |
| Institución: | Universitat Politècnica de València (UPV) |
| Repositorio: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| Idioma: | inglés |
| OAI Identifier: | oai:riunet.upv.es:10251/90926 |
| Acceso en línea: | https://riunet.upv.es/handle/10251/90926 |
| Access Level: | acceso abierto |
| Palabra clave: | Perceived Quality Ordered Probit Urban Public Transport Revealed Preference Survey Focus Group |
| Sumario: | [EN] In this article, an Ordered Logit model is proposed considering systematic and random variations in tastes. The methodology followed for the creation of this model consisted, in first place, in obtaining data using a revelled preferences survey. In the survey, each user had to evaluate, following a qualitative scale, each one of the attributes of the analysed transport system. The variables evaluated in the survey had been grouped into six groups, and for each group, users had to order the attributes belonging to the group, using a ranking based method, from the most important to de least important, and, in the same way, with the groups itself. Once the database is formed, a generic model have been created, establishing this model as a comparative base for the rest. Next, two more models have been estimated one considering systematic users variations and the other one combining the systematic variations with weighted variables. Additionally, three new models have been calculated as an evolution of the previous ones using random variables as representation of systematic and random variations in user’s tastes. The results shows that as model’s complexity increase, an improvement in model fit is achieved. |
|---|