Deploying machine learning methods for health monitoring and cardiovascular disease prediction

Abstract

Smart healthcare has increased to meet the needs of the growing human population and medical expenses. People are all hurrying to catch up with work schedules, academic appointments, and social engagements, especially in this jet age. These often happen at the detriment of our health. Healthcare services, especially those that provide optimal healthcare delivery, face many problems, such as the ineffective provision ofhealth monitoring applications andless emphasis on disease prediction systems.Severalresearchstudies have been carried out in an attempt to proffer solutions to the peculiar problems; however, the problem persists. Therefore, this paper develops a Cardiovascular disease prediction system with specific objectives: implement data analysis for disease prediction using a k-Nearest Neighbors (k-NN)-based machine learning system; evaluate the performance of the developed cardiovascular disease prediction system with existing health monitoring systems. The k-Nearest Neighbors was utilized using a 1025 dataset and 18 attributes collected from the UCI machine learning repository. The results show that k-NN achieved an accuracy of 99.21%. k-Nearest Neighbors algorithm is a non-parametric machine learning that majority voting to classify new case of cardiovascular disease, and non-sensitive to noise and outlier. The proposed model is higher than the existing system, which shows an average accuracy result of 84.63%. The developed machine learning approach will guide healthcare practitioners on the use of machine learning for cardiovascular disease diagnosis and prediction.