Principal component analysis: the basic building block of chemometrics

Principal component analysis: the basic building block of chemometrics The chemometrics is a discipline involving analysis of data and analytical chemistry. It brings together and develops a set of mathematical tools used to extract information from structured and interpretable chemical data. Many definitions of chemometrics have been proposed, but all have common goal such as finding new tools and new ways to exploit information contained in the data to create knowledge. Some of these tools are more and more applied today to metabolomic and/or metabonomic data. They are designed to show what it would be impossible to do with a univariate data analysis. This contribution aims to present one of the basic techniques of the chemometrics: principal component analysis (PCA). After a historical introduction and the basic principles of the technique, a practical example of use of PCA is developed. The interpretation of results is shown with a pedagogical concern, to better show the power of the tool. The data analyzed in this paper concern a kinetic study of the thermal oxidation of edible oil monitored by proton NMR spectroscopy.