Share Email Print
cover

Proceedings Paper

Determination of instrumental precision requirement for expected glucose prediction accuracy
Author(s): Qingbo Li; Xiaotang Hu; Kexin Xu
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Noninvasive measurement of blood glucose is a focal topic in the biomedical engineering field, but no success has been achieved completely as yet. The necessary instrumental precision (i.e., instrumental repeatability signal-to-noise ratio) is prerequisite for the desired prediction accuracy, which have not been investigated systematically. The mathematic relations regarding instrumental precision, regression method and prediction accuracy are put forward. They are tested and analyzed through the determination of glucose concentration in an aqueous solution with near-infrared (NIR) spectroscopy and chemometrics. The results show that the mathematic relations provide a estimate method of prediction accuracy for spectroscopic measurement. In addition, they can determine the necessary instrumental requirement for the expected prediction accuracy when certain regression method is used. This provides theoretic guides to the measuring methods design and references to the prediction accuracy analysis of the noninvasive glucose measurement and even the near-infrared spectroscopy.

Paper Details

Date Published: 12 September 2002
PDF: 8 pages
Proc. SPIE 4916, Optics in Health Care and Biomedical Optics: Diagnostics and Treatment, (12 September 2002); doi: 10.1117/12.483008
Show Author Affiliations
Qingbo Li, Tianjin Univ. (China)
Xiaotang Hu, Tianjin Univ. (China)
Kexin Xu, Tianjin Univ. (China)


Published in SPIE Proceedings Vol. 4916:
Optics in Health Care and Biomedical Optics: Diagnostics and Treatment
Britton Chance; Mingzhe Chen; Gilwon Yoon, Editor(s)

© SPIE. Terms of Use
Back to Top