Terminology for Multivariate Analysis in Surface Chemical Analysis

As a result of its history and scope, multivariate analysis is laden with confusing terminology, with different names given to similar or equivalent terms depending on the technique and the field of application. In close consultation with international experts, the definitions for 23 common terms in multivariate analysis have been developed for the next version of ISO 18115 Surface Chemical Analysis - Vocabulary.

Download PDF file Terminology For Multivariate Analysis 

These definitions continue to be refined with new terms added. Experts in the user community are greatly encouraged to become involved and can do so by contacting the authors.

Practical Chemometrics Guide for SIMS

Multivariate analysis is very powerful in the analysis of SIMS images. Here we provide a basic tutorial in the use of chemometrics, in particular multivariate analysis, and its application to the interpretation of static SIMS spectra and images.

SSIMS datasets contain a vast amount of information and pose huge challenges for data interpretation. As a result, chemometrics methods such as principal component analysis (PCA), multivariate curve resolution (MCR) and partial least squares (PLS) are often employed in data analysis. The benefits of chemometrics are that:

• It provides an objective and statistically reliable approach, eliminates the need for manual peak selection in analysis, and therefore reduces the potential for bias
• It takes into account all the information available in the dataset, giving an improved signal to noise ratio
• It can be fast and automated

However, despite widespread studies of multivariate analysis of SIMS data in literature, the uptake of the techniques into industrial analysis has been slow. For many scientists, there are many ambiguities in the method, confusion in terminology and jargon, which gives low confidence in the results. There is a general need for an improved understanding of basic and practical aspects.

The PDF file Practical Chemometrics Guide 2010 ( 6 MB) provides a basic tutorial, recently given at SIMS Europe 2010, in the use of chemometrics methods. Applications of the techniques to the interpretation of static SIMS spectra and images will be demonstrated using recent examples from literature. The tutorial aims to de-mystify the concepts and terminology, provide guidance on data pre-processing and present a route-map to select the most appropriate methods.

Contents

1. Introduction
2. Linear algebra
3. Factor analysis
   • Principal component analysis
   • Multivariate curve resolution
4. Multivariate regression
   • Multiple linear regression
   • Principal component regression
   • Partial least squares regression
5. Classification
   • Principal component discriminant function analysis
   • Partial least squares discriminant analysis
6. Conclusion