Content

Chemometrics encompasses all mathematical and statistical techniques for the analysis of chemical data. In the first part of this course it will be explained how basic principles (taught in other modules) can be used in chemometrical practice. In the second part more advanced methods will be introduced. These include techniques for multiway data analysis and pure component analysis. Throughout the course, concepts will be illustrated by means of examples from analytical chemistry, process monitoring, environmental studies, food analysis, pharmaceutical and clinical chemistry, etc.

The following aspects of chemometrics will be handled in this course:

• Classical modelling concepts for quantitative calibration: Classical Least Squares (CLS), Inverse Least Squares (ILS), Multivariate Linear Regression (MLR), Principle Component Regression (PCR) and Partial Least Squares Regression (PLSR).
• Necessary steps for the creation and succesful deployment of calibrations; Selection of calibration standards and assesment of the reliability of the models: (Test set validation vs. Cross-validation, model statistics). Special attention will be given to the methods for the selection of the number of principle components or latent variables in the projection methods.
• Methods for data pre-processing and linearization with special attention for the phenomena of light scattering and instrument drift and the methods to deal with these phenomena: derivatives, standard normal variate (SNV), multiplicative signal correction (MSC) and extended multiplicative signal correction (EMSC), external parameter orthogonalization (EPO), generalized least squares weighting (GLSW).
• Variable selection in a chemometric context and some commonly used methods for this such as VIP scores, jack-knifing, uninformative variable elimination, interval PLS and Genetic Algorithm PLS.
• Qualitative analysis in a chemometric context: clustering, discrimination and classification (LDA/QDA, SIMCA, kNN, PLSDA, SVM)
• Nonlinear multivariate calibration/classification methods (kernel PLS, SVR).
• Pure component identification (EFA, EWFA, MCR, O-PLS) with applications on mixture and batch reaction data
• Multi-way methods (MPCA, PARAFAC, N-PLS)

After each chapter, a take home assignment will be given involving application of the taught methods to a given dataset. 

Course material

-H. Martens, T. Naes. "Multivariate calibration", Wiley: Chichester; 1989.
-T. Naes, T. Isaksson, A.M.C. Davies, T. Fearn. "A user-friendly guide to multivariate calibration and classification", IMPublications: Chichester, UK; 2002.
- D.L. Massart, B.G.M. Vandeginste, L.M.C. Buyens, S. De Jong, P.J. Lewi, J. Smeyers-Verbeke, “Handbook of Chemometrics and Qualimetrics”, Elsevier, Amsterdam, 1997.

Format: more information

Starting from the examples discussed in the course, the student applies the theory to a number of manageable cases and summarizes the results in a report.

Is also included in other courses

• G0B70A : Chemometrics