A survey on Machine Learning Methods used for Classification of Cardiovascular disease using ECG and Image Processing Techniques

This research explores critical research gaps in cardiovascular disease detection and prediction, emphasizing the urgent need for further investigation. The identified gaps encompass interpretability issues in machine learning models, insufficient comparisons between contemporary and traditional diagnostic methods, and a lack of attention to generalization performance across diverse patient populations. The review underscores the importance of addressing gaps in model interpretability, real-world challenges of Convolutional Neural Network algorithms, and the exploration of optimal feature sets for predictive modelling. Recognizing limitations such as small datasets and the opaque nature of certain models, the imperative for future studies to prioritize explainable AI frameworks and address broader challenges in data overfitting, cost-effectiveness, and model bias is emphasized. The research advocates for the exploration of alternative algorithms to enhance accuracy and interpretability in cardiovascular disease detection and prediction.