Machine learning techniques for heart disease datasets: A survey

Heart Failure (HF) has been proven one of the leading causes of death that is why an accurate and timely prediction of HF risks is extremely essential. Clinical methods, for instance, angiography is the best and most effective way of diagnosing HF, however, studies show that it is not only costly but has side effects as well. Lately, machine learning techniques have been used for the stated purpose. This survey paper aims to present a systematic literature review based on 35 journal articles published since 2012, where state of the art machine learning classification techniques have been implemented on heart disease datasets. This study critically analyzes the selected papers and finds gaps in the existing literature and is assistive for researchers who intend to apply machine learning in medical domains, particularly on heart disease datasets. The survey finds out that the most popular classification techniques are Support Vector Machine, Neural Networks, and ensemble classifiers.