Premium
Application of procrustean methods to mid‐and near‐infrared spectral data
Author(s) -
Vigneau E.,
Devaux M. F.,
Safar M.
Publication year - 1995
Publication title -
journal of chemometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.47
H-Index - 92
eISSN - 1099-128X
pISSN - 0886-9383
DOI - 10.1002/cem.1180090204
Subject(s) - matrix (chemical analysis) , transformation (genetics) , infrared , calibration , infrared spectroscopy , spectral line , transformation matrix , principal component analysis , computer science , mathematics , chemistry , artificial intelligence , statistics , optics , physics , chromatography , biochemistry , organic chemistry , kinematics , classical mechanics , astronomy , gene
Procrustean methods allow the fitting of a given matrix to another given matrix observed on the same objects. In the traditional approach orthogonal constraints are imposed upon the transformation matrix, whereas in the alternative approach Procrustean analysis may be performed without such constraints. The two methods (with and without constraints) were compared on data dealing with mid‐ and near‐infrared spectra of oil. The aim was to reconstruct the mid‐infrared spectral information using data from the near‐infrared spectra. Unconstrained Procrustean analysis proved to be the more efficient for both the calibration and verification sets. Furthermore, the analysis of the transformation matrix between the two infrared ranges made it possible to indicate wavelengths and wave numbers corresponding to the same chemical groups.