The Prediction of Hypertension Based on Convolution Neural Network

摘要

Hypertension is an illness that often leads to severe and life threatening diseases if left untreated, and earlier diagnosis of hypertension saves enormous lives. In this paper, an optimization scheme for hypertension prediction based on convolution neural network (CNN) is proposed. In order to adapt to the independence of physiological feature data, the convolution kernel was changed to 1D size when CNN was introduced. This method is compared with the following six machine learning techniques: KNN, J48, Random Forest, SMO, Native Bayes, Logistics, with the dataset from the US complex physiological signal research resource website PhsioNet. The results show that the proposed method shows the best performance ( ACC = 89.95%) and is superior to all other machine learning methods tested in this study.