An Intelligent Diagnosis System for Predicting Coronary Heart Disease

The incidence and mortality rates of cardiovascular disease are considered a major concern in the healthcare industry nowadays and remarkable efforts have been made to reduce its effects and mortality rate. Accordingly, an ensemble-based deep learning model is proposed in this paper to predict the risk of cardiovascular disease based on various clinical and laboratory features of individuals. The proposed model includes the combination of convolutional neural networks, long short-term memory networks, extreme gradient boosting, and k-nearest neighbors to increase the accuracy and quality of heart disease prediction. To validate the efficiency of the proposed model, the Cleveland dataset with 303 samples besides a combined dataset including 1190 samples were used in our experiments. Based on the empirical results, the proposed model with an accuracy of 90.59% on the combined dataset has the highest performance compared to the existing model which clearly demonstrates the superiority of our proposed model in cardiovascular disease prediction.