Learning objectives
Learning objectives are the following:
knowledge and understanding
Different approaches for data analysis will be discussed, including univsriate data techinques as well as multivariate approaches. Theory, case studies and practical exercises using data analysis software will be presented. The aim is to make the students understand the pronciples and logic behind the selection of a proper approach for analyzing a specific dataset. The students will be introduced to the experimental design for process optimization for both chemical analysis as well as material synthesis.
Applying knowledge and understanding: the student will be trained for data analysis using software freely available or with institutional licence. In addition, the student will be able to recognize the variables significantly impacting on industrial processes, synthesis and analytical methods. Finally he will learn how to optimize these variables suggesting proper solutions to improve yield, extraction performances, productivity,...
Independent judgement: the students will be able to critically evaluate their expertise. Data analysis and critical evaluation of the obtained results will be required to provide statistically significant conclusions to specific problems. The understanding of the limits of the applied procedure and solution will be evaluated.
Communication skills: the students will be able to discuss about specific chemical/physical problems and the related issues. The student will be able to explain the reasons behind the selection of a specific approach for analyzing the data. Software and presentation could be use for this purpose. The students will be encouraged to discussed with the teacher and the colleagues about problems and how to propose a solution, focusing on multidisciplinary studies typical of material science.
learning ability: the students will understand the logic behind a specific chemometric approach, its limits and assets. They will also use the investigated approaches to solve new problems and analyze the data in a critical way. They will be also able to obtain information about chemometric approaches in literature as well as investigated repositories for obtaining online available datasets. Finally the students will be able to investigate a problem with the aim to optimize the conditions for improving its performances based on experimental design
Prerequisites
- - -
Course unit content
Understand the logic behind statistical analysis, explaining the theory, giving real world examples, with practical applications. During the course different exercises will be proposed. methods based on univariate and multivariate data analysis will be presented. Experimental design for the optimization of analytical conditions as well as industrial processes and reaction condition will be discussed with particular focus on material science
Full programme
recap of the following tests:
- Student t-test
- Variance homogeneity (homoscedasticity)
- F test
Univariate statistics:
- one way ANOVA, two-way ANOVA, two-way ANOVA with interaction, post-hoc test pairwise comparison (Bonferroni)
- interaction among factors
Linear regression
- estimation of coefficient by least square method, coefficient significance;
- the coefficient of determination R squared
- variable correlation: Pearson correlation
experimental design for optimization purposes:
- analysis of the experimental domain
- experimental planning
- full factorial design, fractional design, Plackett-Burmann screening design, experimental error evaluation, curvature test, desirability function
Multivariate data analysis
- introduction to unsupervised and supervised methods
- principal component analysis
- partial least squared regression
Bibliography
George E. P. Box, William Gordon Hunter, J. Stuart Hunter"Statistics for experimenters: an introduction to design, data analysis, and model building" Wiley
James N. Miller, Jane C. Miller, Statistics and Chemometrics
for Analytical Chemistry, Pearson
Teaching methods
The course includes theory of data analysis methods, case studies and exercises. Frontal lessons will be held and discussion of problematic and advantages of the methods will be included. The discussions will also involve the different approaches that can be use to solve a specific case or for data analysis of specific datasets. During the lessons the professor will use slides and exercises and real cases will be performed using freely available software (CAT) or softwares under the university license. The students will be taught to the basics of the used programs. The different steps for data analysis will be analyzed to learn how to perform data analysis of datasets indipendently on their own PCs. The participation to the practical exercise is mandatory to take the exam.
Assessment methods and criteria
The exam will be an oral session of about 30-45 minutes to verify: i) the acquired knowledge on the investigated approaches;
ii) data analysis capabilities using the programs; iii) critical interpretation of the obtained results; iv) experimental planning for condition optimization using experimental design
The evaluation will be given immediately and will be formalized on Esse3 platform.
Other information
- - -
2030 agenda goals for sustainable development
- - -
