Abstract
A method of comparing two spectra using linear and quadratic regression is described. For every wavelength, the log(1/R) value of the first spectrum is used as x, that of the second as y. The correlation coefficient (r) is used to compute a similarity index which can be used to test for identity. The method has the advantage of comparing whole spectra, rather than only certain peaks, while reducing the data to a single variable. For crystalline solids with different particle sizes, better results are obtained if a quadratic rather than a linear equation is used. This procedure allows the successful identification of isomers and sugars, even when there is considerable variation in particle size.
© 1996 NIR Publications
PDF Article
More Like This
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