Abstract
The feasibility of near infrared (NIR) spectroscopy and least squares support vector machine (LSSVM) for rapid determination of reducing sugar content in potatoes was investigated. Statistical models were developed between reducing sugar content and NIR spectra by LSSVM. Compared to partial least squares regression (PLSR), the performance of LSSVM model was a slight superior with higher correlation coefficient of prediction (rp) of 0.992 and lower standard error of prediction (SEP) of 0.130%. The results show that LSSVM is a suitable tool to deal with nonlinear problem, and NIR technique can be used for measuring reducing sugar content in potatoes rapidly.
© 2014 Optical Society of America
PDF ArticleMore Like This
Dirk Grosenick, Oliver Steinkellner, Heidrun Wabnitz, Rainer Macdonald, Thoralf Niendorf, Kathleen Cantow, Bert Flemming, Karen Arakelyan, and Erdmann Seeliger
BS3A.80 Biomedical Optics (BIOMED) 2014
Nathan. R. Newbury, Esther Baumann, Christopher Cromer, Ian Coddington, Fabrizio Giorgetta, Greg Rieker, Laura Sinclair, William Swann, and Alex Zolot
LW3D.1 Laser Applications to Chemical, Security and Environmental Analysis (LACSEA) 2014
Reember Cano, Luis C. Tapia, Gabriela Castellano, and Rickson C. Mesquita
BM3A.32 Biomedical Optics (BIOMED) 2014