Partial Least-Squares Methods for Spectral Analyses. 1. Relation to Other Quantitative Calibration Methods and the Extraction of Qualitative Information

Abstract

Partial least-squares (PLS) methods for spectral analyses are related to other multivariate calibration methods such as classical least-squares (CLS), Inverse least-squares (ILS), and principal component regression (PCR) methods which have been used often In quantitative spectral analyses. the PLS method which analyzes one chemical component at a time Is presented, and the basis for each step In the algorithm is explained. PLS calibration Is shown to be composed of a series of simplified CLS and ILS steps. This detailed understanding of the PLS algorithm has helped to Identify how chemically interpretable qualitative spectral information can be obtained from the Intermediate steps of the PLS algorithm. These methods for extracting qualitative information are demonstrated by use of simulated spectral data. the qualitative Information directly available from the PLS analysis is superior to that obtained from PCR but Is not as complete as that which can be generated during CLS analyses. Methods are presented for selecting optimal numbers of loading vectors for both the PLS and PCR models In order to optimize the model while simultaneously reducing the potential for overfitting the calibration data. Outlier detection and methods to evaluate the statistical significance of results obtained from the different calibration methods applied to the same spectral data are also discussed. © 1988, American Chemical Society. All rights reserved.