Inaccuracies in forecasting create serious problems for transportation projects. To improve forecasting accuracy, researchers must investigate the reasons for the inaccuracies. However, very little research has been done, and the percentage of errors for each reason has rarely been examined. After perusing various studies on forecasting inaccuracies, the authors determined that an in-depth case study was necessary. This paper describes such a case study: an overestimate in a forecast of travel demand, forecast using the four-step method, for the Tokadai Line rail service in a suburb of Nagoya, Japan. The paper examines the following factors behind the errors: target area, population, modal split, failure to consider the effects of a competing railway, inappropriate selection of modal choice model, and an incomplete network. Through this case, three implications of general inaccuracies in forecasting are presented. First, while previous studies have shown that input uncertainties are larger than model uncertainties, this does not mean that inputs cause larger errors. Second, the uncertainties related to inputs must be prioritised. Finally, the reasons for the inaccuracies are more difficult than expected to identify, since the same inaccuracies can be explained by more than one reason. © 2013 Elsevier Ltd.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below