心拍変動時系列による血圧値推定に対するニューラルネットワークの適用

概要

論文の詳細を見る
The purpose of this study is to measure the blood pressure for a long time continuously and non-invasively. In this paper, it is proposed that the method of estimation of the systolic blood pressure (SBP) from the heart rate variability. Our method is to estimate SBP from the time series of the R-R intervals using the neural network. The R-R intervals related to SBP are detected using regression analysis as the preprocessing of the neural network application. In addition, the neural network according to time zone and the individual is constructed. Moreover, the learning process of the neural network is evaluated using the correlation coefficients between the estimation value and the measurement value. As a result, the correlation coefficients between the estimation value and the measurement value were 0.67(n=2596) or less.
大同工業大学の論文