THE APPLICATION OF DISAGGREGATE CHOICE MODELS TO TRAVEL DEMAND FORECASTING: ISSUES AND METHODS

This paper describes three important issues in the application of disaggregate models for aggregate prediction, and develops approaches to resolving these issues. First, is the identification of aggregation procedures which can be used with disaggregate models to produce aggregate predictions of travel demand. Second, are procedures to select an appropriate disaggregate model from either the application context or other environments. Third, is a procedure to simplify the requirements for prediction of exogenous variables when the choice process is amenable to description by the multinomial logit or nested logit models. (TRRL)