270275 VO Chemometrics and Data Analysis in Multidimensional Analysis (2016W)

Modern analytical methods for continuously monitoring e.g. industrial processes or environments usually need to simultaneously acquire a range of analytes. This leads to large amount of data in comparably short time. Their evaluation thus usually needs computer assisted methods as they are developed by chemometrics, whose main fields are:
- Patter Recognition: (hierarchical) cluster analysis, principal component analysis
- Modelling of Data: regression methods in one and more variables, principal component analysis, neural networks.
- Time Series Analysis: autocorrelation functions
- Quality Assurance and Good Laboratory Practice
- Experimental Design.
In addition to simultaneously analyzing multicomponent mixtures, chemometrics also allows to classify samples according to not directly quantifiable criteria, such as discriminating different wines from each other by their smell or taste with so-called "artificial noses" or "artificial tongues". The main focus of the lecture will be on the actual analytical application of these techniques. Therefore, in-deep mathematical derivations will be foregone as far as possible.