A Novel Two-Stage Heart Arrhythmia Ensemble Classifier

Atrial fibrillation (AF) and ventricular arrhythmia (Arr) are among the most common fatal cardiac arrhythmias in world. Electrocardiogram (ECG) data, collected as part of UK Biobank, represents an opportunity for analysis classification these two diseases UK. The main objective our study is to investigate a two-stage model individuals with AF Arr Biobank dataset. current literature addresses heart very extensively. However, data used by researchers lack enough instances diseases. Moreover, proposing separation normal abnormal cases, we have improved performance classifiers detection each specific disease. Our approach consists stages classification. In first stage, features ECG input classified into classes: abnormal. At second further categorised Arr. A diverse set such QRS duration, PR interval RR interval, well covariates sex, BMI, age other factors, modelling process. For both stages, use XGBoost Classifier algorithm. healthy population present has been undersampled tackle class imbalance data. This technique applied evaluated using dataset from UKBioBank taken at rest repository. results paper follows: proposed measured F1 score, Sensitivity (Recall) Specificity (Precision). system 87.22%, 88.55% 85.95%, average Score, sensitivity specificity, respectively. Contribution significance: level indicates that automatic participants more precise efficient if done manner. Automatic this way would mean early diagnosis prevention serious consequences later their lives.