FSHOADC - Feature Selection and Hyperparameter Optimization Based Arrhythmia Detection and Classification: A Machine Learning Approach for Arrhythmia Detection using ECG Signals

One essential and often used tool for the diagnosis of cardiovascular disorders is the electrocardiogram (ECG). With the emergence of artificial intelligence (AI), it is possible to have learning-based approaches for the automatic detection of several types of heart diseases. Despite the fact that deep learning models have shown remarkable results in automated ECG analysis, they frequently lack model interpretability, which is crucial for healthcare applications. Certain techniques made use of ECG data, which can help identify a variety of illnesses, especially arrhythmias. Precisely categorizing heartbeats is necessary for arrhythmia detection, which is an essential part of ECG analysis. The ECG signals used in this study are pre-processed for heartbeat segmentation and noise reduction. Therefore, it is evident that reliable automated detection methods are required. The proposed framework is subjected to multiple optimizations in terms of preprocessing, feature engineering, and hyperparameter tuning. In our study, we created a brand-new method for detecting arrhythmias in ECG signals termed FSHOADC. The experimental results show that the recommended machine learning framework and optimizations might enhance the performance of arrhythmia diagnosis and detection. It is also noted that the suggested method outperforms other works in the literature, achieving an accuracy of 98.8%.