The transferability of disaggregate travel demand models is viewed as a problem of predicting the values of model parameters in a new context by using the estimated values from an existing model. The approach tested in this paper is to use survey data from the new context to estimate transfer-scale parameters that correct for context-specific effects. The empirical tests investigate a variety of transfer models with different numbers of transfer-scale parameters ranging from complete reestimation of the entire model specification to none at all. The results support the partial transfer model in which scale parameters are estimated for a limited number of functional subgroups of variables such as alternative specific constants, level-of-service attributes, and personal characteristics.
Samenvatting