MULTIVARIATED DATA ANALYSIS IN ENVIRONMENTAL MATRICES

The course aims to introduce students to chemometrics tools that can be useful in studying and interpreting complex systems.
The course will start with the description of the multivariate structure of data for the characterization of a chemical/environmental system. The main methods covered in the course will be: Pattern Recognition, with particular reference to Cluster Analysis, Principal Component Analysis (PCA) and the development of predictive multivariate models (PCR and PLS) by means of QSAR (Quantitative Structure-Activity Relationship) strategy and Experimental Design.
Particular emphasis will be given to the development and the validation of multivariate calibration models, and to the study and the understanding of practical problems concerning environmental systems. Different cases of study will be discussed in-depth.
These knowledge are extremely important inside the Degree’s program, because the students will acquire the necessary skills for interpreting environmental data and understanding the mechanisms and the dynamics of environmental processes in chemical, ecological, geological, hydrogeological and biological field. That is the starting point for implementing strategies of remediation, monitoring and evaluation of environmental biodiversity, as well as defense of natural resources.
The course has a predominant experimental approach and is focused on the understanding and the practical use of the main methods of Pattern Recognition.

The student will be able to acquire:
1) the appropriate skills for using chemometrics methods of multivariate analysis, such as Pattern Recognition methods and Experimental Design;
2) the ability to apply these methods for solving new environmental problems, by being able to interpret new data related to environmental systems. These skills will be given by the presentation of specific cases of study in which chemometrics methods are applied;
3) a sufficient critical approach on choosing the right chemometrics methods for solving environmental problems and on extrapolating from the results useful information, which can lead him to implement a decisional strategy of action on an environmental system.
The student will have to learn a consistent and correct use of the language and terminology, so he will be able to propose and communicate this methodological approach even to people who don’t have these skills and with whom he will have to collaborate in a professional future.

High school basic knowledge of mathematics

Pattern Recognition.
Preliminary data treatment, classification and clustering methods: K-NN, Cluster analysis.
Principal Component Analysis (PCA): theory, use, applications; the SIMCA method.
Multivariate correlation models: MRA, PCR and PLS methods, validation criteria, strategy for the selection of the best model dimension and its optimization.
Experimental Design.
Factorial Design, D-efficiency and D-optimal design. Practical use of chemometrics software.
A part of the course will be dedicated to the study of real cases from the relevant literature.

Didactic material and presentations provided by the professor.
Roberto Todeschini: "Introduzione Alla Chemiometria". EDiSES, Napoli.
D.L. Massart et al: "Chemometrics:a Textbook", Data Handling in Science and Technology, 2, ELSEVIER, Amsterdam.

The final mark will be given after an oral examination concerning the topics discussed during the course. The student will be asked to discuss cases of study and show the ability to apply the chemometrics approach to real environmental problems.

Classroom lessons. The lessons will have the main goal of explaining the “philosophy” of the chemometrics approach, which is completely different from the one commonly used in the traditional scientific method.

Italian

STRUCTURE AND CONTENT OF THE COURSE COULD CHANGE AS A RESULT OF THE COVID-19 EPIDEMIC.

oral