|
The Chemometrics course teaches the student to program, apply and interpret results from the most commonly used modern chemical data analysis techniques. These data visualization and analysis techniques are crucial in contemporary analytical chemistry, biology and medicine.
At the end of the course the student:
- ...can select the proper data analysis method, based on the chemical/biochemical question and the principles behind each method.
- ...can write MATLAB programs to executte these methods on several real life data sets of different size and complexity.
- ...can correctly interpret and validate the method results, to translate them to answers to the chemical/biochemical question.
The chemometric techniques that the student will master are listed under Subjects.
|
|
|
During the statistics course, the main focus was on univariate chemical and biological data analysis. However, most measurements on chemical and biological systems are multivariate. Examples are (N)IR or NMR spectra, mass spectra and chromatographic data. Analysis of such data requires multivairate methods, which provide better predictions and more insight into the chemical system; better calibration curves, improved disease predictions, earlier detection of process errors etc. This course teaches you to program and interpret the main tools for multivariate data visualization and analysis.
The course gives an overview of the basic chemometric methods, sometimes described as 'Chemical Data Science'. Examples from all fields of the natural sciences will make this course useful for students with very diverse backgrounds. A short and non-exhaustive list of areas where these techniques are applicable is: organic and industrial chemistry, proteomics and metabolomics (finding biomarkers that are indicative for specific diseases) and the detection of unknown pollutants in drinking water. During the course, the various chemometric methods are studied and exercised thoroughly by programming in MATLAB. The student will also learn how to interpret the resulting models to obtain information for further study.
The techniques that will be covered are :
- Principal Component Analysis: Exploratory Analysis of multivariate data
- Hierarchical Clustering
- Linear Discriminant Analysis for Classification
- Statistical model validation
- Partial Least Squares Regression and Discriminant Analysis
This course will involve programming in MATLAB; Lectures are shared with Chemometrics for Molecular Life Sciences and practicals are course-specific. The course (or Chemometrics for MLS) is obligatory for a master internship at the Department of Analytical Chemistry.
Instructional Modes
Subjects start with an introductory lecture, which is followed by a computer course in which the relevant methods are programmed in MATLAB. The student will use the programmed method to do a data analysis of a chemical experiment, which may be discussed and completed in a second computer course.
|
 |
|
|
- Statistics (NWI-MOL028) or DATA: statistics and programming (NWI-MOL150)
- Basic Linear Algebra
- Basic knowledge of MATLAB (MATLAB course or DATA) This is a course in the theme 'Methods'.
|
|
|
Three assignments: first two in pairs of students, third should be done individually. All grades must be sufficient and the third assignment counts for 50% in the final grade.
|
|
|
A follow-up course is Pattern Recognition in the Natural Sciences (SM299). The course will be given in parallel with the course Chemometrics for MLW (MOL109), that focuses more on interpretation of the model results. Students may switch between MOL109 and this course until after the first computer course.
|
|