Chemometrics II – Regression and Partial Least Squares (PLS)

Course Description

If Principal Components Analysis (PCA) is the most important machine learning/chemometric method, then Partial Least Squares (PLS) regression is a very close second. Chemometrics II — Regression and PLS covers regression methods starting with Classical Least Squares (CLS) and Multiple Linear Regression (MLR) and culminates in Principal Components Regression (PCR) and PLS Regression. Students will learn to safely apply the methods to create predictive models in a variety of applications. 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, MATLAB for Machine Learning and Chemometrics and Chemometrics I — PCA, or equivalent experience.

Course Outline

1. Nomenclature and Conventions
   
2. Classical Least Squares (CLS)
   
3. Inverse Least Squares (ILS) models 
   
4. Multiple Linear Regression (MLR)
   
5. Ridge Regression (RR)
   
6. Principal Components Regression (PCR)
   
7. Determination of number of PCs – Cross Validation
   
8. Partial Least Squares (PLS)
   
9. Interpreting PLS Models
   
10. Outlier detection and model diagnostics
    
11. Example datasets: pseudo-gasoline, SFCM temperature/level, styrene-butadiene copolymers