Modeling attitude to risk in human decision processes: An application of fuzzy measures

Several models of the human decision process have been proposed, classical examples of which are utility theory and prospect theory. In recent times, the theory of fuzzy measures and integrals has emerged as an alternative meriting further investigation. Specifically, one is interested in the degrees of disjunction and conjunction and the veto and favor indices that represent the tolerance measure of the decision maker. Though several theoretical expositions have appeared in contemporary literature, empirical studies applying these concepts to the real world are scarce. This paper reports two studies based on a model of strategic telecommunication investment decisions from a research work involving a survey of executives. The first study involves building fuzzy models corresponding to each individual decision maker with the results grouped based on the decision makers’ propensity to risk as determined by their degrees of disjunction. The Shapley indices and the interaction effects are determined for each pooled dataset corresponding to each group. To contrast this approach with those of conventional nomothetic comparisons of decision policies, the decision makers are grouped based on a clustering analysis of the individual linear regression models. The data for each cluster are pooled and the fuzzy measures learned from the dataset are analyzed for comparison purposes. The results not only serve as a demonstration of fuzzy measure analysis as a viable approach to studying qualitative decision making but also provide useful methodological insights into applying fuzzy measures to strategic investment decisions under risk.