You are viewing a javascript disabled version of the site. Please enable Javascript for this site to function properly.
Go to headerGo to navigationGo to searchGo to contentsGo to footer
In content section. Select this link to jump to navigation

Research on the multiple linear regression in non-invasive blood glucose measurement

Issue title: Frontiers in Biomedical Engineering and Biotechnology – Proceedings of the 4th International Conference on Biomedical Engineering and Biotechnology, 18–21 August 2015, Shanghai, China

Guest editors: Feng Liu, Dong-Hoon Lee, Ricardo Lagoa and Sandeep Kumar

Article type: Research Article

Authors: Zhu, Jianminga; b; * | Chen, Zhenchengb; *

Affiliations: [a] School of Life and Environmental Sciences, Guilin University of Electronic Technology, Guilin 541004, China | [b] Guangxi Experiment Center of Information Science, Guilin 541004, China

Correspondence: [*] Address for correspondence: Jianming Zhu, School of Life and Environmental Sciences, Guilin University of Electronic Technology, No. 1, Jinji Road, Guilin City, Guangxi Prov., PA 541004, China. Tel.: +86 0773-2293135; Fax: +86 07732293135; E-mail: zjmcsu@126.com. Zhencheng Chen, Guangxi Experiment Center of Information Science, No. 1, Jinji Road, Guilin City, Guangxi Prov., PA 541004, China. Tel.: +86 0773-2292232; Fax: +86 0773-2292232; E-mail: chenzhcheng@163.com.

Keywords: Non-invasive blood glucose measurement, metabolic energy conservation, multiple linear regression, Clarke error gird, sensor

DOI: 10.3233/BME-151334

Journal: Bio-Medical Materials and Engineering, vol. 26, no. s1, pp. S447-S453, 2015

Published: 2015
Get PDF open access

A non-invasive blood glucose measurement sensor and the data process algorithm based on the metabolic energy conservation (MEC) method are presented in this paper. The physiological parameters of human fingertip can be measured by various sensing modalities, and blood glucose value can be evaluated with the physiological parameters by the multiple linear regression analysis. Five methods such as enter, remove, forward, backward and stepwise in multiple linear regression were compared, and the backward method had the best performance. The best correlation coefficient was 0.876 with the standard error of the estimate 0.534, and the significance was 0.012 (sig. <0.05), which indicated the regression equation was valid. The Clarke error grid analysis was performed to compare the MEC method with the hexokinase method, using 200 data points. The correlation coefficient R was 0.867 and all of the points were located in Zone A and Zone B, which shows the MEC method provides a feasible and valid way for non-invasive blood glucose measurement.

Show more