Chemometrics

Recommended Prerequisites

Not applicable.

Teaching Methods

- Interactive lectures given by the teacher(s).

- Practical, computational, problem solving.     

Learning Outcomes

Acquire proficiency in methods, algorithms and tools in multivariate analysis, for comprehensive chemical applications.

Work Placement(s)

No

Syllabus

1. Introduction

Data processing and information handling. Chemometrics and related disciplines.

2. Basic Statistics

3. Experimental design

Simple factorial design. Two level procedures. Fractional replication. Composite designs.

4. Methods of correlation and time series.

Covariance, correlation and regression. Time series: autocovariance and autocorrelation.

5. Cluster analysis

Definition. Distance and similarity. Hierarchical techniques. Partitioning-optimization techniques.

6. Principal component and factor analysis

Interpretation. Principal components in two and n dimensions

Analysis of factors: determination and rotation

7. Supervised pattern recognition

Decision rules. Linear discriminant analysis. The method of k-nearest neighbors.  Decision trees. Bagging and random forests.

8. Advanced regression

Principal component regression. Partial least squares. Structure-activity relationships. 

9. Neural networks

Algorithms. Hidden layers and deep learning.

Head Lecturer(s)

Alberto António Caria Canelas Pais

Assessment Methods

Assessment
Resolution Problems: 25.0%
Exam: 75.0%

Bibliography

L. Sachs, “Applied Statistics: a Handbook of Techniques”, 2^nd Ed., Springer, New York, 1984.

D.L. Massart, B.G.M. Vandeginste, S.N. Deming, Y. Michotte e L. Kaufman, “Chemometrics: a text book”, Elsevier, Amsterdam, 1988.

R.G. Brereton, Chemometrics : Data Analysis for the Laboratory and Chemical Plant, Wiley (Chichester, 2003).