Project Description

This project focuses on classifying arrhythmia, an irregular heartbeat condition, using the MIT-BIH Arrhythmia dataset. Various supervised learning models, including Naive Bayes, Decision Tree, 1D-CNN, and Deep Neural Networks, were employed. The key challenge addressed was class imbalance, with 75% of the data representing normal heartbeats. Data augmentation and reduction techniques were applied for fairness. 1D-CNN proved most effective, achieving high accuracy and recall in classifying arrhythmia types, offering a robust solution for real-time ECG signal analysis.