An efficient feature reduction approach for arrhythmia disease detection utilizing SVM

Abstract

Arrhythmia is viewed as a hazardous infection causing genuine medical problems in patients, when left untreated. An early determination of arrhythmias would be useful in saving lives. This examination is led to group patients into one of the sixteen subclasses, among which one class addresses nonappearance of sickness and the other fifteen classes address electrocardiogram records of different subtypes of arrhythmias. The exploration is done on the dataset taken from the University of California at Irvine Machine Learning Data Repository. The dataset contains a huge volume of highlight measurements which are decreased utilizing SVM-RFE based component determination strategy. The dataset contains a huge list of capabilities which is diminished utilizing an improved component choice strategy named as covering technique. The proposed covering technique is based on a SVM-RFE calculation to choose the main highlights from the given dataset. The chose subset of highlights at that point goes through a preprocessing step to present a consistency in the dispersion of information. Since help vector machine (SVM) is perceived to have the advantage of giving an eminent execution in grouping stage.