E. Corchado et al. (Eds.): HAIS 2012, Part II, LNCS 7209, pp. 369–380, 2012. © Springer-Verlag Berlin Heidelberg 2012 From Likelihood Uncertainty to Fuzziness: A Possibility-Based Approach for Building Clinical DSSs Marco Pota, Massimo Esposito, and Giuseppe De Pietro Institute for High Performance Computing and Networking ICAR-CNR via P. Castellino 111, Naples 80131, Italy {marco.pota,massimo.esposito,giuseppe.depietro}@na.icar.cnr.it Abstract. For data classification, in fields like medicine, where vague concepts have to be considered, and where, at the same time, intelligible rules are required, research agrees on utility of fuzzy logic. In this ambit, if statistical information about the problem is known, or can be extracted from data, it can be used to define fuzzy sets and rules. Statistical knowledge can be acquired in terms of probability distributions or likelihood functions. Here, an approach is proposed for the transformation of likelihood functions into fuzzy sets, which considers possibility measure, and different methods arising from this approach are presented. By using real data, a comparison among different methods is performed, based on the analysis of transformation properties and resulting fuzzy sets characteristics. Finally, the best method to be used in the context of clinical decision support systems (DSSs) is chosen. Keywords: DSS, Fuzzy sets, fuzzy logic, likelihood, probability, possibility, transformation. 1 Introduction Decision-making, especially in fields where vague concepts have to be handled, represents a very challenging issue. The main purpose of recent research efforts, in medicine, economy and automated processes control, is the classification of data items in a finite number of conclusions, in other words, the determination of the membership of an object to a specific class. An instrument skilled to help user’s decision-making is called Decision Support System (DSS). Using data acquired from actual cases, whose membership to a class is known, a certain amount of knowledge can be acquired, about the relationships between data items and respective class membership. If this knowledge is properly modeled, new incoming data items can be classified. In knowledge-based DSSs, data are processed and modeled by exploiting a rule representation formalism, while in data-driven DSSs, they are used as training set for statistical and machine learning models. A number of particular approaches exists, consisting in various modules of both knowledge-based and data-driven DSSs, and hybrid combination of them.