The Study on Separability Criteria Suitable for Cardiac Data

There is a need to develop a criterion to measure the class separability of cardiac data before the decision-making logic is applied. This article explores the use of separability criteria to measure the separability of cardiac data. Then a new separability criterion was presented, which was constructed by combining standard deviation with Euclidean center distance. The criterion was applied to extract features of cardiac arrhythmias, measure the performance of the features, and build the decision tree for multiclass classification. The data in this study were obtained from MIT-BIH database. The experimental results show that it is an effective and practical separability measurement criterion.