An Overview of Atomic Spectrometric Techniques
Classical Linear Regression by the Least Squares Method
Implementing a Robust Methodology: Experimental Designs and Optimisation
Ordinary Multiple Linear Regression and Principal Components Regression
Partial Least‐Squares Regression
Multivariate Regression using Artificial Neural Networks and Support Vector Machines
About this book
This is the first book for atomic spectroscopists to present the basic principles of experimental designs, optimization and multivariate regression. Multivariate regression is a valuable statistical method for handling complex problems (such as spectral and chemical interferences) which arise during atomic spectrometry. However, the technique is underused as most spectroscopists do not have time to study the often complex literature on the subject. This practical introduction uses conceptual explanations and worked examples to give readers a clear understanding of the technique. Mathematics is kept to a minimum but, when required, is kept at a basic level. Practical considerations, interpretations and troubleshooting are emphasized and literature surveys are included to guide the reader to further work. The same dataset is used for all chapters dealing with calibration to demonstrate the differences between the different methodologies. Readers will learn how to handle spectral and chemical interferences in atomic spectrometry in a new, more efficient and cost-effective way.
Jose Andrade-Garda is based in the Department of Analytical Chemistry at the University of A Coru±a where he specializes in quality control and chemometrics. Within the field of chemometrics, his main interests are multivariate regression and pattern recognition methods. In the atomic spectrometry arena, he has applied formal optimization techniques to optimize analytical protocols and used multivariate regression tools to cope with spectral and chemical interferences in ETAAS.