Discriminant Analysis and Class Modelling of Spectroscopic Data

Kemsley, E. K.

ISBN 10: 0471978302 ISBN 13: 9780471978305

Publisher: Wiley–Blackwell, 1998

Synopsis

Discriminant analysis is the generic name given to statistical techniques used to analyze a large amount of experimental data. This package includes the software, Win–Das which was developed by the author to analyze spectroscopic data, and is already being used successfully.

About this title

From the Back Cover

This unique book and software package ("Win–DAS") is both an introduction to discriminant analysis and class modelling, and a powerful tool for analysis of the user?s own spectro8copic data. The software includes worked examples based on real data sets, including distinguishing between different types of coffee, and species identification in fruit pulps. By reading the book and working through these tutorials, the user will be become familiar with the following chemometrics methods.

• principal component analysis (PCA)
• partial least squares (PLS) for discriminant analysis
• linear discriminant analysis (LDA) and canonical variate analysis (CVA)
• the class modelling methods of UNEQ (UNEQual dispersed classes) and SIMCA (Soft Independent Modelling of Class Analogy).

The software also allows users to input their own data and can be installed on any PC connected to a spectrometer for easy and direct transfer of files. This versatile software package can be used in any situation which requires analysis of a large number of spectra, with many examples in the tutorials taken from the author?s own field of expertise, food science. System requirements for the Win–DAS software are : CD–ROM drive, Microsoft Windows 3.1 or 95, a mouse compatible with Windows 3.1 or 95, a colour monitor with VGA (640x480) or SVGA (800x600) resolution, and approximately 5Mb of available hard drive space. A spreadsheeting and graphing package such as Microsoft Excel is also useful.

Edition details

• Publisher: Wiley–Blackwell
• Publication date: 1998
• Language: English
• ISBN 10: 0471978302
• ISBN 13: 9780471978305
• Binding: Software
• Number of pages: 150