A Soft Computing Paradigm For A Medical Data Mining Tool To Predict Risk of Coronary Heart Events

Abstract

In this paper, a Fuzzy-Classified Neural learning soft computing tool (FCNL) is proposed for predicting the intensity of risk in Coronary Heart occurrences. The presented model utilizes medical data collected from clinical findings on cardiac patients. The concept of decision trees is employed to classify the attributes that add to the Coronary Artery Disease (CAD). The output obtained as a result is then transformed to crisp if-then rules and then fuzzified into a database of fuzzy rules. A fuzzy-classified neural learning method based on supervised learning is exercised to enhance fuzzy membership functions. The performance and efficiency of the new medical data mining system, in terms of accuracy of prediction is presented against the real-life data.