Análise de demanda por transportes de passageiros via modelos de regressão georeferenciados

• Autor(a) principal: Ribeiro, Valéria da Cruz
• Data de Publicação: 2012
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/3930
• topic: Demand for Transport; Spatial Regression; Weighted Regression Geographically; Spatial Statistics; Demanda por Transportes; Regressão Espacial; Regressão Ponderada Geograficamente; Estatística Espacial; Engenharia Civil; Transportes; Análise de regressão; Análise espacial (Estatística); 624

This dissertation - Analysis of Demand for Passenger Transport via Regression Models georeferenced - presents, and a methodology for the construction of spatial regression models and geographically weighted, a risk assessment when compared to traditional regression models and regression models with dummies variables in order to forecast demand for travel to the city of Vitoria, capital of Espirito Santo, in order to obtain information that can subsidize the transportation planning more effectively. For this, we used data from the household survey of origin and destination (OD) held in 1998 in the metropolitan region of Vitoria, four models were calibrated regression modeling of travel demand: Traditional Model Regression, Regression Model dummy Regression Model Space and Geographically Weighted Regression Model. After calibration, the models were tested from the application data in the household survey of origin and destination conducted in 2007 in the same city, to compare and validate the estimate. We conclude that the main hypothesis, or part thereof, considered in this work was confirmed that a regression model spatial or geographically weighted distances can be more explanatory than conventional regression models, since the calibration of travel demand models by weighted regression model showed values of statistical adjustments smaller than the other models.