Using principal components analysis for dimension reduction prior to compositional analysis

Publication Name
COMPSTAT 2004 Proceedings
Publisher
Physica-Verlag
ISBN
3790815543
Abstract
We investigate the use of principal components analysis (PCA) as an aid to compositional analysis. We study a data set concerning measurements of chemical markers (n-alkanes and alcohols) on artificially-created mixtures of grass and clover shoots. We go on to apply our procedure to mixture data sets based on actual Fourier Transform Infrared (FTIR) spectroscopy measurements on soil, where the very large number of markers recorded means that some form of dimension reduction is a necessity.
Year
2004
Category
Book Chapter
Output Tags
pre-2006 Soil & Environmental Science