Abstract (EN):
The multivariate statistical analysis of cis and trans isomers of fatty acid profiles of eight margarine brands obtained by HRGC/FID/capillary column was carried out based on biplots applied to principal component analysis (PCA) and canonical correlation analysis (CCA). It is shown that while predictive biplots are the best choice for interpretation purposes, interpolative biplots are very useful for classification of new observations that were not used for the construction of the principal component or canonical dimension axes. Copyright (C) 2004 John Wiley Sons, Ltd.
Idioma:
Inglês
Tipo (Avaliação Docente):
Científica
Nº de páginas:
9