Chemometrics I — Principal Components Analysis (PCA)

Course Description

Chemometrics I — PCA, concentrates on what is perhaps the most important method in machine learning and chemometrics, Principal Components Analysis. PCA can be used for exploratory data analysis, pattern recognition, data prescreening, and is part of many other methods such as SIMCA sample classification. It is also used for preprocessing and data compression in a wide variety of applications such as Support Vector Machines (SVMs) and Artificial Neural Networks (ANNs). This course covers the basics of PCA in depth, concentrating on interpretation of PCA models. The course includes hands-on computer time for participants to work example problems using PLS_Toolbox or Solo.

Prerequisites

Linear Algebra for Machine Learning and Chemometrics and MATLAB for Machine Learning and Chemometrics or equivalent experience.

Course Outline

1. Nomenclature and conventions
   
2. Data transformation-Linearization
   
3. Data centering and scaling
   
4. The PCA decomposition
   
5. Interpreting scores and loadings plots
   
6. Q and T² statistics
   
7. Outliers
   
8. Determination of number of factors to keep
   
9. Example datasets: Wine, Octene, Arch, and Olive Oil