Modeling the Dynamic Individual Activity Decision Behavior Through the Influence Diagram Model

Abstract

Using the Influence Diagram (IFD), this paper attempted to model the human decision behaviors on choosing their daily activities by considering both the features of human decisions and the expected utilities outcomes of decisions. The 2004 time use survey data from the National Statistical Office (NSO) which represents 2,573 individuals in Bangkok were utilized as the case study in the paper. As the results, there are 19 proposed IFD models based on types of populations and number of members in the household. Results from the model validation processes also presented that all 19 IFD models were successfully simu-lated the activities (ACT) and the decisions of whom in the household to do activities with (WHO). The IFD model is expected to be better represent individual activity decisions than the Classification and Regression Tree (CART) model and the Bayesian network model as its incorporated the utility scores in the decision processes.