An Extensive Survey on Heart Disease Prediction

Abstract

Prediction of occurrences of heart diseases an important work in medical field. A growing number of studies use various methods and datasets for heart disease prediction. Still, a substantial part of my research work lacks categorization and systematization. Hence, there is a necessity to take stock of current knowledge in this field. In this sense, this paper carries out a review of the literature on heart disease prediction. Analysis of a total of 150 papers addressing heart disease prediction using Data Mining (DM), Machine Learning (ML) and Deep Learning (DL) methods published between 2010 and 2020 was carried out. The purpose and the contribution of this paper is to provide a clearer picture of the subfields in heart disease prediction by concentrating on two aspects. First, a recent research to categorize the main areas of specialization in heart disease prediction is reviewed and the challenge in each category has been addressed. Second, the thematic analysis is carried out to identify a specific method under each category and a suitable method is also recommended. Despite the large number of publications, this present study identifies deep learning methods with unstructured dataset by providing better results. This study helps researchers for understanding the research gap for their future study in heart disease prediction.