A fuzzy classification model for myocardial infarction risk assessment

The use of data mining approaches for analyzing patients trace in different medical databases has become an important research field especially with the evolution of these methods and their contributions in medical decision support. In this paper, we develop a new clinical decision support system (CDSS) to diagnose Coronary Artery Diseases (CAD). According to CAD experts, Angiography is most accurate CAD diagnosis technique. However, it has many aftereffects and is very costly. Existing studies showed that CAD diagnosis requires heterogeneous patients traces from medical history while applying data mining techniques to achieve high accuracy. In this paper, an automatic approach to design CDSS for CAD assessment is proposed. The proposed diagnosis model is based on Random Forest algorithm, C5.0 decision tree algorithm and Fuzzy modeling. It consists of two stages: first, Random Forest algorithm is used to rank the features and a C5.0 decision tree based approach for crisp rule generation is developed. Then, we created the fuzzy inference system. The generation of fuzzy weighted rules is carried out automatically from the previous crisp rules. Moreover, a critical issue about the CDSS is that some values of the features are missing in most cases. A new method to deal with the problem of missing data, which allows evaluating the similarity despite the missing information, was proposed. Finally, experimental results underscore very promising classification accuracy of 90.50% while optimizing training time using UCI (the University of California at Irvine) heart diseases datasets compared to the previously reported results.