Simultaneous Multivariate Analysis of Multiple Data Matrices

The application of pattern recognition methods to two-way tables of data emanating from mass spectrometers, chromatographs, and a variety of other complex analytical instruments is now widespread. In particular principal components analysis (Pearson,(1) but see Gnanadesikan(2) for a more recent introduction) is used to obtain low-dimensional representations of the relationships between rows and columns of two-way tables. In pyrolysis mass spectrometry this technique, usually called “factor analysis,” has been particularly powerful.(3,4)