Basic Chemometrics

Course Description

Basic Chemometrics focuses on what are perhaps the two most important chemometric methods, Principal Components Analysis (PCA) and Partial Least Squares (PLS) regression. PCA can be used for exploratory data analysis, pattern recognition and data prescreening/cleaning. PCA is part of many other methods and is also used for preprocessing data in a wide variety of applications (e.g. SVMs and ANNs). This course covers the basics of PCA, concentrating on interpretation of PCA models. The course continues with the motivation for regression models. Multiple Linear Regression (MLR) is introduced along with Principal Components Regression (PCR). Advantages and problems with these methods are discussed, and it is shown how Partial Least Squares (PLS) mitigates these issues. Examples of using PLS in spectroscopic calibrations are demonstrated. The course includes hands-on computer time for participants to work example problems using PLS_Toolbox or Solo.

Prerequisites

Linear Algebra for Chemometricians and MATLAB for Chemometricians or equivalent experience.

Course Outline

  1. Nomenclature and Conventions
  2. Data Centering and Scaling
  3. The Principal Components Analysis (PCA) Decomposition
  4. Interpreting Scores and Loadings
  5. Residual Q and Leverage T^2 Statistics
  6. Outliers
  7. Determination of Number of PCs
  8. Additional PCA Examples and Hands-on Exercises
  9. Motivation for Regression
  10. Multiple Linear Regression (MLR)
  11. Principal Components Regression (PCR)
  12. Partial Least Squares Regression (PLS)
  13. Determination of Number of PCs or LVs
  14. Interpreting PLS Models
  15. Outlier Detection and Model Diagnostics
  16. Additional Examples and Hands-on Exercises