Computational process modeling has been introduced as an alternative approach to utility-maximizing framework to deal with the complexity of activity-based models of travel demand. ALBATROSS, a rule-based system, used data mining algorithms to derive choice rules underlying activity-travel patterns. In the context of a project that attempts to better include household as opposed to individual decision making into the original model, this paper describes the results for the car allocation decisions. The CHAID algorithm is applied to derive a decision tree for the car allocation decisions in automobile deficient households using a large activity diary data set recently collected in the Netherlands. The results show a satisfactory improvement in goodness of fit of the decision tree model compared to the null model. The probability of the male getting the car is considerably higher than the female getting the car in many condition settings. In only 16% of the condition settings, the female has the highest probability of getting the car. Accessibility of the work location by car relative to slow mode appears to be the most influential factor when both male and female work. For the covering abstract see ITRD E137145.
Abstract