Classification of ecg signals analysis using rough sets theory and a new classi̇fi̇cati̇on algori̇thm approach (FWRSC)

Early diagnosis and treatment of heart diseases such as heart attack is very important. ECG signals, for the diagnosis of various diseases and disorders of the heart is the most effective method used for early diagnosis. Because the incidence of incomplete and inconsistent data sets in the health field, the proper interpretation of the ECG signal is very important. Rough set theory (RST) is a rule-based method used in extracting information and analysis for expert analysis such as fuzzy sets. Rough sets make it suitable by organizing incomplete, inconsistent and uncertain data sets for evaluation. In this study, ECG signals classified with the RST firstly. At this point, using the minimum parameters, which would help doctors for more rapid and early detection of a pattern has been established. This model has been classified the ECG signals with an accuracy close to 85%. Later, a new classification that the weight matrix created scoring system by using RST and similarity-based method has been developed and it's been named as Feature Weighted Rough Set Classification (FWRSC). This method is compared with the classification method in WEKA for 5 difference data sets. The experimental results showed that FWRSC showed higher performance than many methods in WEKA. In addition, in terms of accuracy, it showed the highest performance with the overall average of 67.47 for 5 difference data sets.