Premium
Multivariate data analysis and bivariate regression studies applied to comparison of two multi‐elemental methods for analysing wine samples
Author(s) -
Sagrado S.,
PérezJordan M. Y.,
Pastor A.,
Salvador A.,
de la Guardia M.
Publication year - 2002
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.723
Subject(s) - principal component analysis , analyte , multivariate statistics , bivariate analysis , statistics , data matrix , partial least squares regression , mathematics , principal component regression , calibration , data set , residual , chemometrics , matrix (chemical analysis) , analytical chemistry (journal) , chemistry , chromatography , algorithm , clade , biochemistry , gene , phylogenetic tree
Two inductively coupled plasma mass spectrometry (ICP‐MS) methods which permit multi‐elemental analysis in wine samples have been compared following two strategies. First, a multivariate tool based on principal component analysis (PCA) was employed for a global (all analytes) qualitative comparison of the two methods. A single plot based on the confidence limits of the Q and T 2 PCA model statistics corresponding to the ‘standard’ method results (calibration set) was used to check the comparability of the ‘candidate’ method (test samples). The residual matrix (after test matrix interpolation into the PCA model) gives qualitative information about the nature of the main errors. This approach is compatible with the presence of missing data. Second, comparison of the methods, analyte by analyte, based on bivariate least squares (BLS) regression was used. This approach, which uses the uncertainty information from both methods and performs a joint test for the intercept and slope parameters, permits appropriate comparison when the two methods have non‐constant standard deviations. Copyright © 2002 John Wiley & Sons, Ltd.