Predicting the presence of serious coronary artery disease based on 24 hour Holter ECG monitoring

The purpose of this study was to evaluate the usefulness of classification methods in recognizing a cardiovascular pathology. Based on clinical and electrocardiographic (ECG) Holter data we propose a method for predicting a coronary stenosis demanding revascularization in patients with a diagnosis of a stable coronary heart disease. A possible solution of this problem has been set in a context of rough set theory and methods. The rough set theory introduced by Zdzisław Pawlak during the early 1980s provides a foundation for the construction of classifiers. From the rough set perspective, classifiers presented in the paper are based on a decision tree calculated on a basis of a local discretization method, related to the problem of reducts computation. We present a new modification of a tree building method which emphasizes the discernibility of objects belonging to decision classes indicated by human experts. The presented method may be used to assess the need for the coronary revascularization. The paper includes results of experiments that have been performed on medical data obtained from Second Department of Internal Medicine, Collegium Medicum, Jagiellonian University, Kraków, Poland.