In the laboratories advanced instrumental techniques (such as NIR spectroscopy and UV_Vis ..) are currently employed not only with analytical purposes but also for on-line monitoring of production processes. These modern instruments are capable of performing analysis in a short time, but they provide a huge amount of data that can be processed only with the use of appropriate chemometric techniques. This course aims to provide the basis to extract useful information from multivariate data.
The aim of the course is to provide the students with simple and powerful tools for basic multivariate data analysis. Among the possible applications of multivariate statistical analysis to chemical data, it will be shown how multivariate quality control can detect "bad" samples (i.e., not complying with the product specifications).
All lessons are composed of a first theoretical part where the theory of chemometrics (without going into much detail of mathematical algorithms) is explained and a part of computer exercises. In this second part, a specific problem and the related data set are described to the students, that must try to extract the desired information using a statistical software.
Exploratory data analysis (Principal Component Analysis) to visualize the data structure; classification methods to identify a sample as belonging to one or more groups of previously-classified samples; regression methods to determine the amount of a component, property, or other value based on the measured X-block variables. PCA diagnostics.
Ricevimento: Every day, by appointment.
ELEONORA RUSSO (President)
BRUNO TASSO (President)
The exam takes place at the computer. Students are provided with a data set and they have to extract the information useful for the problem by using different techniques of multivariate analysis (PCA, LDA, SIMCA ...) implemented on a statistical software used during the lessons.
