Abstract
A more precise estimate of the accuracy of near infrared (NIR) spectroscopy is obtained when the measured standard errors of cross validation (SECV) and prediction (SEP) are corrected for imprecision of the reference data. The significance of correction increases with increasing imprecision of reference data. Very high precision of reference data obtained through replicate analyses under reproducibility conditions may not be the optimal goal for the development of calibration equations. In a situation of limited resources, the precision of the reference data should be related to the obtainable accuracy of the spectroscopic system. Investigation of several routine applications based on the partial least-squares (PLS) regression technique showed that increased precision of calibration data only resulted in marginal improvements in true accuracy if the total standard error of reference results from the beginning was less than the estimated true accuracy of the corresponding NIR calibration.
© 2002 NIR Publications
PDF Article
More Like This
Sensitivity of MRI-guided near-infrared spectroscopy clinical breast exam data and its impact on diagnostic performance
Michael A. Mastanduno, Junqing Xu, Fadi El-Ghussein, Shudong Jiang, Hong Yin, Yan Zhao, Kelly E. Michaelsen, Ke Wang, Fang Ren, Brian W. Pogue, and Keith D. Paulsen
Biomed. Opt. Express 5(9) 3103-3115 (2014)
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription