A Study on Coronary Disease Prediction Using Boosting-based Ensemble Machine Learning Approaches

Abstract

In today’s world, a gigantic measure of information is generated in the medicinal services industry. In most cases, this information is underutilized and is not constantly made use to the full degree. Utilizing this gigantic measure of information, certain types of disease can be identified, anticipated or even restored. These diseases e.g. cardiovascular disease, malignant growth of cancer cells, tumor or Alzheimer’s disease can cause an enormous risk to mankind. In this paper, we attempt to focus on coronary heart disease prediction. Utilizing the Machine Learning (ML) approaches, the coronary disease can be anticipated. The medicinal information, for example, Blood Pressure (BP), hypertension, diabetes, the number of cigarettes smoked every day, etc. can cause coronary disease and they are taken as input and afterward, these data are used to forecast the possibility of occurring this disease for oneself. This model would then be able to be utilized to foresee future therapeutic information. Several boosting algorithms of ensemble techniques like Adaptive Boosting (AdaBoost), Extreme Gradient Boosting (XGBoost), Gradient Boosting Machine (GBM), Light GBM (LGBM), and Category Boosting (CatBoost). The accuracy of the model utilizing every one of these algorithms is determined. Their performance is shown using several other parameters. At that point, the one with a decent exactness is taken as the model for foreseeing the coronary disease.