During the past two decades, a significant amount of effort has been put into improving the models with a special focus on the (theoretical) model structure. Recently, a great deal of the research has been directed towards error component models where differences between individuals have been modelled by means of statistical distributions. Differences between individuals have been realised some time ago, and have generally been incorporated by means of segmentation of the data. Basically, segmentation is a simple division of observations by e.g. the purpose of the trip. Segmentation implies that a separate model is estimated for each segment. The aim of the segmentation has been to improve the accuracy of the models (reproduction of observed data) by minimising the within segment variation of the preferences - preferably by use of segments that would be applicable from a planning or strategic point of view. A description follows of how segmentations affect traffic models. Then a description of a method that can improve model fit by means of segmentation. Finally, an application of the method.
Abstract