In this paper we present the Decision Support Framework (DSF) of the NOESIS platform. NOESIS addresses wide scale integration and visual representation of medical intelligence in cardiology and aims at the development of a Web-based personalized system with enhanced intelligence that supports health professionals in taking the best possible decision for prevention, diagnosis, and treatment. The core of the NOESIS project is a set of Fuzzy Expert Systems (FES), one for each cardiovascular sub-domain, automatically generated from the DSF. An initial set of crisp rules, generated using data mining techniques, is employed to define a fuzzy model, using the sigmoid function and fuzzy equivalents of the binary operators. The parameters used to define the models are tuned using a global optimization algorithm