Multivariate Analysis of Quality

Data analysis is a vital part of science today, and in assessing quality, multivariate analysis is often necessary in order to avoid loss of essential information. Martens provides a powerful and versatile methodology that enables researchers to design their investigations and analyse data effectively and safely, without the need for formal statistical training. * Offers an introductory explanation of multivariate analysis by graphical 'soft modelling' * Minimises mathematics, providing all technical details in the appendix * Presents itself in an accessible style with cartoons, self-assessment questions and a wide range of practical examples * Demonstrates the methodology for various types of quality assessment, ranging from human quality perception via industrial quality monitoring to environmental quality and its molecular basis All data sets available FREE online on "Chemometrics World" (http://www.wiley.co.uk/wileychi/chemometrics)

Preface. Acknowledgements. OVERVIEW. Why Multivariate Data Analysis? Qualimetrics for Determining Quality. A Layman's Guide to Multivariate Data Analysis. METHODOLOGY. Some Estimation Concepts. Analysis of One Data Table X: Principle Component Analysis. Analysis of Two Data Tables X and Y: Partial Least Squares Regression (PLSR). Example of Multivariate Calibration Project. Interpretation of Many Types of Data X and Y: Exploring Relationships in Interdisciplinary Data Sets. Classification and Discrimination X_1, X_2, X_3: Handling Heterogeneous Sample Sets. Validation X and Y. Experimental Planning Y and X. APPLICATIONS. Multivariate Calibration: Quality Determination of Wheat From High-speed NIR Spectra. Analysis of Questionnaire Data: What Determines Quality of the Working Environment? Analysis of a Heterogeneous Sample Set: Predicting Toxicity From Quantum Chemistry. Multivariate Statistical Process Control: Quality Monitoring of a Sugar Production Process. Design and Analysis of Controlled Experiments: Reducing Loss of Quality in Stored Food. Appendix A1: How the Present Book Relates to Some Mathematical Modelling Traditions in Science. Appendix A2: Sensory Science. Appendix A3.1: Bi-linear Modelling Has Many Applications. Appendix A3.2: Common Problems and Pitfalls in Soft Modelling. Appendix A4: Mathematical Details. Appendix A5: PCA Details. Appendix A6: PLS Regression Details. Appendix A7: Modelling the Unknown. Appendix A8: Non-linearity and Weighting. Appendix A9: Classification and Outlier Detection. Appendix A10: Cross-validation Details. Appendix A11: Power Estimation Details. Appendix A12: What Makes NIR Data So Information-rich? Appendix A13: Consequences of the Working Environment Survey. Appendix A14: Details of the Molecule Class Models. Appendix A15: Forecasting the Future. Appendix A16: Significance Testing with Cross-validation vs. ANOVA. References. Index.