Three-Dimensional Category Analysis Using Probabilistic Approach

Abstract

Category analysis is a method to find the average response of trips made by a particular category of analysis unit such as a person or household. Conventional analysis is done by household categories whereby the causal household socioeconomic characteristics that are supposed to influence trip generations are cross-classifeid in a two-dimensional plane, and the household trip rates are calculated. A modified approach that would take care of not only the household socioeconomic characteristics but also an individual's travel attribute characteristics has been taken up in this study. The analysis has been extended from the household plane to the individual's travel attribute plane. The anomalies pertaining to sparse or no data due to stratification of household socioeconomic variables and travel attributes such as trip length have been taken into account in this study. A probabilistic approach to model formulation has been attempted. Calibration and validation of the model have also been undertaken.