Fuzzy techniques of pattern recognition in risk and claim classification

Applications of fuzzy set theory to property-liability and life insurance have emerged in the last few years through the work of Lemaire (1990), Cummins and Derrig (1993, 1994), and Ostaszewski (1993). This article continues that line of research by providing an overview of fuzzy pattern recognition techniques and using them in clustering for risk and claims classification. The classic clustering problem of grouping towns into rating territories (DuMouchel, 1983; Conger, 1987) is revisited using these fuzzy methods and 1987 through 1990 Massachusetts automobile insurance data. The new problem of classifying claims in terms of suspected fraud is also addressed using these same fuzzy methods and data drawn from a study of 1989 bodily injury liability claims in Massachusetts.