A systematic review and analysis of the heart disease prediction methodology

Most of the decisions in medical diagnosis are taken on the basis of experts' opinions. In the case of heart diseases, however, the experts' decisions do not always reach a consensus since the pattern of heart disorders varies considerably among patients. Researchers have been making continuous efforts to detect heart diseases at the primary stages by using different methodologies in order to increase the chances of curing a condition that has one of the highest mortality rates in the world. The three main objectives of this study were to analyze the global impact of heart diseases on the basis of mortality rates, to assess the risk of heart diseases in different age groups, and to discuss the advantages and disadvantages of methodologies that have been used previously for predicting heart disease at the primary stage. The mortality rate due to heart diseases was assessed according to attributes such as age, population group, clinical risk factors, and geographical locations. Different methodologies were analyzed on the basis of results obtained from literature searches in IEEE, Elsevier, Springer, and other publications. The percentage of deaths due to heart diseases increase with age, indicating that the risk of developing heart disease is directly proportional to age. The analysis of various methodological approaches indicated that data mining and the combination of optimization methods can be effective in predicting heart disease at the initial stages. The current data available on heart diseases can help design better frameworks for predicting new cases. The statistics of heart disease-related death shows a worrying trend world-wide. This study concludes that a framework based on hybrid approaches consisting of the combination of classification and clustering methods of data mining, along with biological system inspired algorithms, can prove to be a landmark in the field of heart disease prediction and detection.