Mode Choice Models Transferability: A Bayesian Update for Two Cities in Iran

Bayesian updating is introduced as a method to improve results of model transferability intended for decreasing cost of related studies. The main goal of this paper is to study the effects of Bayesian updating on transferability of aggregate and disaggregate mode choice models. These models are calibrated for a city in Iran and transferred to another city. Results of employing transferability measures such as Transfer Index (TI) and goodness of fit of transfer model are evaluated in two forms of naïve and Bayesian updating. Results show that Bayesian updating can noticeably enhance transferability of disaggregate models, but not as much in aggregate models. It is concluded that applying Bayesian update approach is reliable only when transferability of naïve model is not rejected. 