Skip to main content
ARS Home » Research » Publications at this Location » Publication #81443
Title: DISCRIMINANT ANALYSIS OF FOOD INGREDIENTS BY MID-INFRARED DIFFUSE REFLECTANCE SPECTROSCOPY

Author
item Reeves Iii, James
item ZAPF, CHARLES - MC CORMICK SPICE COMP,INC

Submitted to: Spectroscopy Conference Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: 8/31/1997
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: The objective of this research was to determine if mid-infrared diffuse reflectance spectroscopy could be used to discriminate among various food ingredients. One hundred and six samples consisting of 15 milk powders, 20 hard&soft wheat flours, 14 cheese powders, 10 cheese seasonings, 21 ranch seasoning and 22 onion dressing powders were scanned as ground neat powders susing diffuse reflectance on a Fourier transform spectrometer equipped wit a custom made sample transport device. Samples were scanned (64 co-added scans per samples) at 4 and 16 cm-resolutions from 4000 to 400 cm with KBr used as the background sample. Samples were also scanned (32 co-added scans per sample) in the near-infrared on a scanning monochromator from 1100 to 2498 nm (9091 to 4003 cm) using a rotating sample cup. Every third sample was set aside as part of a validation set of 33 samples with the remaining 73 samples used to develop the discriminant models. Discriminant analysis was performed using "Mahalanobis Distance by Principal Component Analysis with Residuals" using the "Averaged Predicted Distance" indicator to determine the number of factors, obtained from a one-out cross validation analysis, to use for each grouping. Spectra were pretreated using mean centering and multiplicative scatter correction. Results based on 4 cm-resolution mid-infrared spectra were as accurate as those developed using the near-infrared spectra, with all 33 validation samples correctly classified. These results demonstrate that mid-infrared spectra of neat samples of food products can be used for product discrimination. The ability to use mid-infrared spectra for such product identification could be very useful when a mid-infrared spectrometer is already available or is needed for other reasons.