Transferability of Freight Trip Generation Models

Abstract

The main objectives of this paper are to assess and define ways to enhance the transferability of freight trip generation (FTG) models. After the key premises that should guide the development of FTG models have been presented, the paper assesses transferability in two ways. The first is through analyses of how well representative FTG models are able to estimate the actual FTG for a number of external validation cases. The second is through FTG econometric models that assess the statistical significance of binary variables that represent specific geographic locations. In addition, the paper introduces and assesses the accuracy of a synthetic correction procedure that is intended to improve the transfer-ability and quality of the estimates provided by the FTG rates available in the literature. The results show that the models developed as part of the National Cooperative Freight Research Program's Project 25, Freight Trip Generation and Land Use, have better prediction capabilities than the models included in other compilations. In addition, the synthetic correction procedures improve transferability, and no locational effects are present in the test data.