Estimating Unobserved Attributes of Unchosen Alternative Transportation Modes Using the Normal Distribution Probability Density Function

Abstract

In the calibration of the mode choice model, the observed disaggregated data, such as the personal survey data, are theoretically more desirable than the aggregated data, such as the zonal based average or total data. Zonal based aggregate data, however, are commonly used to calibrate a mode choice model because of a lack of observed information on the values of the attributes of a specific traveler in the disaggregated data. Especially, there are usually no observed data on the service attributes of unchosen alternative transportation modes for individual passengers in household travel survey data. This study suggested a method of estimating the values of unobserved attributes of alternative transportation modes using the observed statistical distribution of the attribute's values such as the travel time for each transportation mode for each zone pair. The study also showed a sample application of the method in calibrating the mode choice model using household survey data.