Detection of Cardiac Disease and Association with Family History using Machine Learning

In todays’ era, the health sector has become a rich source of information due to the existence of many data analysis tools with the huge availability of any patients’ disease datasets. Since, Coronary Heart Disease (CHD) is one of the world’s leading causes of death, detecting it can help reduce mortality rates. When making predictions using conventional methods, the complexity of the data and correlations is a problem. The purpose of this research is to leverage historical medical data to predict CHD using machine learning technologies. Random forest (RF), Support vector machine (SVM), and k–Nearest neighbour classification (k–NN) are three supervised learning approaches used in this paper to detect whether a patient is suffering from CHD or not with a higher prediction rate. The authors also validate the associativity of Heart disease and family history(famhist). South African (SA) heart disease dataset of 462 instances with 10–fold cross-validation method is used here for experimental results. RF model is found promising in detecting CHD with 95% accuracy. Moreover, famhist and CHD are found statistically significant associated (p <0.05) with χ
²
(1)=34.27.