Probability and statistics for engineers and scientists (2nd edn), Anthony J. Hayter, Duxbury, Pacific Grove, CA, 2002, ISBN 0-534-38669-5, xii + 916 pp, £33.00

✍ Scribed by Richard G. Brereton

• Book ID: 101833635
• Publisher: John Wiley and Sons
• Year: 2002
• Tongue: English
• Weight: 48 KB
• Volume: 16
• Category: Article
• ISSN: 0886-9383
• DOI: 10.1002/cem.765

✦ Synopsis

Chemometricians come into the subject from many disciplines. A few are formally trained as statisticians, but many have backgrounds in diverse areas such as analytical chemistry, organic chemistry, chemical engineering or biology, among others. The latter groups have often last encountered statistics at school, or as a minor course in university that they have dropped for their favoured discipline, and are returning to statistics perhaps because they have used a software package to analyse data or want to optimize a reaction.

Many scientists and engineers first encountering chemometrics lack an insight into basic statistics. Knowing how to set up a factorial design or to interpret loading plots from a PCA package is useful, but at some stage in the education of a serious chemometrician it is useful to be able to return to basic statistical principles. Hayter's text is a useful addition to the personal libraries of readers of this journal. Although one may in day-to-day practice be analysing data such as by PLS or MLR, at some stage one will come to a point where one needs to understand more basic ideas; for example, how most statistical tests are based on the normal distribution, or analysis of variance, or the basis of linear regression. The pioneers of chemometrics had a good grounding in fundamental statistics, so it is important that the next generation of practitioners likewise has this basis. Most textbooks in chemometrics do not have room for a full course in statistics, and their readers enter from widely differing perspectives.

An important feature of this book is that there are a very large number of numerical problems in each chapter, typically well over 50. The solutions to half the problems in the main body of the chapter are presented at the end of the book, leaving the remaining problems as exercises for students that can be set on courses. There is a very wide range of real data sets coming from different application areas, which are used to illustrate many of the methods, most of which are available to download on the publisher's website (www.duxbury.com) in a variety of different formats. The book is not tied to any specific computer package, but Minitab is used to illustrate many of the examples.

The first five chapters give a good basic grounding in probability theory, dealing primarily with distributions. Chapter 5, on the normal distribution, is particularly helpful to the chemist and also introduces the lognormal, F-, t-and w 2 distributions.

Chapters 6±10 discuss basic statistical methods, including descriptive statistics, estimations, statistical estimations, confidence intervals, hypothesis testing, comparison of two population means, and finally discrete data analysis, where variables take on a certain defined number of categories (e.g. acceptable and unacceptable quality of a product). These