Fusion of Distribution Diversity Measures to Optimize Cross-media Features for Arrhythmia Prediction by Ensemble Classification

Abstract

Arrhythmia is a common cause of death in people affected by Cardiovascular diseases (CVD). In clinical practice, computer-aided arrhythmia prediction using electrocardiograms is critical, and it has the potential to minimize mortality caused by untrained clinicians. To predict arrhythmia on electrocardiograms (ECGs), machine learning models have been designed based on the ECG signal features architecture, which is a biologically inspired neural network Furthermore, computer-aided approaches frequently succeed in early identification of arrhythmia scope from ECG readings received, which are frequently provided by individuals or medically dispersed networks such as the internet of medical things (MIOT). Distributing computer-assisted therapy techniques that have been successfully used to treat human arrhythmia is all the rage these days, and machine learning is the latest trend in this area. Particularly well-liked in computer-aided arrhythmia prediction technologies are machine learning methodologies. The majority of recent research focuses on the use of cross-media traits in machine learning training. However, false alarms are commonly generated by machine learning models because of the huge dimensionality of the cross-media feature values employed in training. The high dimensional features of cross-media in the learning phase was addressed in this paper, and a fusion strategy was presented to minimize the total data points. It also established a method for predicting arrhythmias from electrocardiograms using the Arrhythmia Prediction by Ensemble Classification approach (APEC). The suggested technique is a classification methodology that selects appropriate cross-media characteristics by combining diversity assessment factors. The suggested method's growth in the prediction accuracy of both labels is the subject of the experimental investigation. The cross-validation statistics of deep genetic ensemble classification (DGEC), and support vector machine (SVM-Ensemble) were compared to modern techniques of machine learning-based arrhythmia prediction algorithms to scale the performance of the APEC.