Prediction of coronary heart disease using novel deep learning algorithm with red deer optimization

Abstract

Early detection of coronary heart disease (CHD) can help decrease these rates because it is one of the leading causes of mortality in the globe. When employing standard approaches to predict, the complexity of the data and relationships provides a problem. In order to predict CHD using Deep Learning (DL) technology, this study will utilize previous medical data.This study proposes a novel Hybrid Recurrent Neural Network (HRNN) with a Red Deer Optimization (RDO) for a prediction of CHD.For this study, CHD datasets are gathered and preprocessed using min-max normalization to standardize the raw data. Independent component analysis (ICA) is used to feature extraction. The proposed HRNN+RDO method’s performance is evaluated and also compared with existing methods in terms ofvariety of metrics to identify the best suitable method regarding CHD prediction. When it came to predicting CHD patients, the proposed model performed with existing methods, with the highest accuracy of 99.75%, the lowest MAE of 37.82%, the lowest RMSE of 35.7%, and the prediction time of 0.25 seconds.