Hierarchical learning model for early prediction of coronary artery atherosclerosis

Artificial intelligence plays an ever-increasing role in developing human-like intelligent machines. In the modern world, physical activities in which people indulge have reduced and this has made them prone to heart diseases such as coronary artery disease (CAD). Coronary artery atherosclerosis (CAA) is one of the main causes of CAD and therefore early prediction of CAA is indispensable to prevent the risk of people getting affected by CAD. This work presents the machine learning model which provides more information on the exceptional cases while retaining the existing traditional classifier for early prediction of CAA. The proposed model performs outliner detection using local outlier factor (LOF) and class balancing using synthetic minority oversampling technique. Genetic algorithm is used for prominent feature selection and utilises support vector machine and neural network as the classifier. Two datasets namely UCI dataset and South African heart disease dataset are used to implement the model. Results show that the proposed model gives better accuracy for the above datasets along with the traditional methods.