Advanced Trip Generation/Attraction Models

• Published in: Procedia, social and behavioral sciences social and behavioral sciences, 2014-12, Vol.160, p.430-439
• Main Authors: Amavi, Alexandre A., Romero, Juan P., Dominguez, Alberto, dell’Olio, Luigi, Ibeas, Angel
• Format: Article
• Language: English
• Subjects: Generation/attraction models; Spatial correlation models; Spatial error; Spatial lag
• ISSN: 1877-0428; 1877-0428
• DOI: 10.1016/j.sbspro.2014.12.155

In this paper, advanced trip generation/attraction models are proposed. A multiple linear regression (MLR) model has been created from zonal data. These models are compared to each other by analyzing their hypothesis and the required adjustments. Additionally, advanced generation/attraction models considering spatial correlation are proposed, and their improvements with reference to previous models not considering spatial correlation are analyzed.A global spatial correlation model that conducts a joint review of every unit in the sample to determine whether the spatial units are randomly distributed or in accordance with a certain pattern is specified. To consider cluster situations in a given area, a local spatial correlation model aimed to measure the spatial autocorrelation to place each observation is defined. The models are applied to the Santander metropolitan area (in Spain) in order to obtain advanced generation/attraction models in that city. For Santander, the models considering spatial dependence among observations have better results than the MLR models.