A Comprehensive Investigation on the Performance of Traditional Machine Learning in Comparison to Deep Learning for Early Cardiovascular Disease Diagnosis

Cardiovascular diseases are one of the leading causes of death, and as such it has become a necessity to find a solution for decreasing the overall numbers. This study is directed toward the study of machine-learning algorithms in comparison with deep-learning algorithms for early diagnosis of cardiovascular disease to identify who is at risk of developing it, and find which algorithms give a more accurate prediction. However, due to the nature of the dataset being small, the deep-learning algorithms are at a disadvantage. This study aims to provide a roadmap to find optimal algorithms for developing ML- or DL-based technologies that help in early detection, which may also aid in the study of the risk factors that contribute to the development of cardiovascular diseases. According to the study, among machine-learning algorithms, the Gaussian Naive Bayes model demonstrated the highest accuracy rate of 88.2%, and among deep learning algorithms, the ANN architecture demonstrated the highest accuracy of 87.2% in test data.