You are here: Home » Study Plan » Subject



master degree course in FOOD SAFETY AND CONTROL

Course year 2
Teaching units Unit unico
Food Technologies (lesson)
  • TAF: Compulsory subjects, characteristic of the class SSD: CHIM/01 CFU: 5
Food Technologies (exercise)
  • TAF: Compulsory subjects, characteristic of the class SSD: CHIM/01 CFU: 3
Exam type oral
Evaluation final vote
Teaching language Lingua italiana
Contents download pdf download


The aim of the course is to furnish theoretical and practical knowledge related to the principal techniques of experimental design and multivariate data analysis, both for exploratory data analysis and to build multivariate calibration and classification models, in order to give to the students the possibility to apply the main chemometric techniques to the analysis of chemical, technological and sensory data measured on food matrices.
For a more complete understanding, please refer to the reading of the expected learning outcomes.

Admission requirements

Knowledge acquired in the first level (Bachelor's) degree in General Chemistry, Analytical Chemistry, Mathematics.
Knowledge of the following topics of basic statistics is also suggested: descriptive statistics, distributions, confidence limits, significance tests, ANOVA, control charts, uni- and multi-linear regression, as well as the use of spreadsheets (e.g. Microsoft Excel) to perform basic statistical data analysis. For those who feel to have not sufficient preparation on these topics, it must be pointed out that in the in the training offer of the Department of Life Sciences is present an optional course of Processing of Experimental Data (Elaborazione dei dati sperimentali).

Course contents

The subdivision of the contents by CFU is indicative. It may in fact undergo changes during the teaching activity in light of the feedback from students.
- Introduction to the topics of the teaching activity: definition of multivariate analysis and of Chemometrics; the multivariate approach to the analysis of data structure and to the extraction of useful information (0.5 CFU).
- Experimental design: 2- and 3-levels factorial designs, central composite design, Dohelert design. Sceening designs: FFD, Plackett-Burmann. Constrained domains: D-Optimal. Mixture designs. Hill climbing techniques: steepest ascent and simplex (1.5 CFU).
- Exercises with MS-Excel on experimental design (1 CFU).
- Exploratory data analysis: structure, variability and information. Clustering: distance and similarity; linkage methods; agglomerative clustering; dendrograms. Column-wise data preprocessing: mean centering, autoscaling. Principal Component Analysis (PCA): geometric interpretation. Scores, loadings, percentage of explained variance. Scree plots, loading and score plots, biplots, Q vs. T^2 plots. PCA on signals; row-wise preprocessing: SNV, smoothing, derivatives, detrending, MSC. Multivariate control charts (Q vs. T^2); contribution plots. Multivariate control charts. Matrix definition of PCA; orthogonality of score vectors and orthonormality of loading vectors; eigenvalues and percentage of explained variance. Examples of real applications of Clustering and of PCA to problems in the agri-food sector (2 CFU).
- Exercises with chemometric software (Eigenvector PLS-Toolbox) on Clustering and PCA.
- Multivariate calibration and classification: from multilinear regression (MLR) to latent variables (LVs) based calibration methods; Principal Component Regression (PCR ) and Partial Least Squares (PLS). Selection of the number of LVs: overfitting and underfitting. Crossvalidation and validation with external test setRMSE e R^2 statitsics; comparison between the parameters calculated in calibration, crossvalidation and external validation. Examples of applications of multivariate calibration models to problems in the agri-food sector. Similarities and differences between calibration and classification; class modeling and discriminant analysis: operating principles of the SIMCA and PLS-DA algorithms. Models and class thresholds. Main statistical parameters for the evaluation of classification models: sensitivity, specificity and efficiency in classification. Examples of applications of multivariate classification models to problems in the agri-food sector (1 CFU).
- Exercises with chemometric analysis software (Eigenvector PLS-Toolbox) on multivariate calibration and classification (1 CFU).

Teaching methods

The teaching activities take place face to face and are in Italian language. Attendance is not compulsory but recommended, especially for computer exercises. Teaching methods include: - classroom lectures (40 hours, 5 CFU), which include guided discussions on aspects relating to design of experiments and multivariate analysis of experimental data, as well as examples of applications of chemometric techniques to problems in the agri-food sector; - computer exercises (24 hours, 3 CFU) on experimental data sets in the agri-food sector using software for statistical data analysis (Microsoft Excel and Eigenvector PLS-Toolbox); the exercises take place in the computer room and include a part led by the lecturer and a part in which the students autonomously process the data and discuss with the lecturer the problems encountered and the results obtained.

Assessment methods

At the end of the course there is an oral exam which also includes a practical test consisting in a computer exercise of experimental data analysis. The exam will take place within the educational calendar of the training offer and the date will be defined by contacting the professor via email. The interview, lasting approximately 40 minutes, will be semi-structured and will start from a request to summarize the topics of the student’s thesis internship in progress or, if this has not yet begun, of the thesis / internship of the student bachelor’s degree. Starting from these topics, a simulated problem will be then posed, for which it is necessary to apply one or more chemometric methods among those described in the course. Once the correct approach has been identified by the student, both the practical and theoretical aspects will then be discussed, in order to evaluate the ability to correctly describe and apply chemometric techniques, to know how to correctly analyze and evaluate the results, and to plan correctly the whole data analysis process. The interview will then be followed by a practical test on the computer lasting about 20 minutes; the software used in the exercises will be applied to an experimental dataset that is as much relevant as possible to the previously discussed issues. The indicators (ascertained characteristics) for the evaluation of the oral exam are: - ability to use and connect knowledge (30%) - ability to discuss and deepen topics (40%) - mastery of scientific language (30%) The final score will progressively reflect the different achievement of the evaluation indicators.

Learning outcomes

Knowledge and understanding:
- describe and discuss the main techniques of experimental design and multivariate data analysis in the specific professional sector of food science and technology;
- illustrate the use of software for chemometric processing of experimental datasets relating to problems in the food sector.
Applied knowledge and understanding:
- independently use some of the most popular software for data processing with multivariate techniques;
- master the main methods of experimental design and multivariate data analysis.
Autonomy of judgment:
- autonomously determine the most appropriate data analysis strategies for solving common problems in the food industry, also by correlating transversal skills relating to the chemical, physical and technological properties of food, process conditions and analytical techniques.
Communication skills:
- clearly communicate the theoretical knowledge acquired;
- use correctly and appropriately the language, concepts and models acquired to discuss the most appropriate strategies for planning the experiments and for data analysis;
- describe and justify the results obtained using the chemometric techniques.
Learning ability:
- acquire the appropriate terminology and methodology for finding information useful for autonomous professional training, with particular attention to IT tools for data processing.


Durante il corso verranno messe a disposizione sulla piattaforma Microsoft Teams le dispense delle lezioni frontali in lingua inglese e videoregistrazioni relative agli argomenti delle lezioni ed alle esercitazioni al computer.

Testi di consultazione:
- J.C.Miller and J.N. Miller "Statistics and Chemometrics for Analytical Chemistry" Ellis Horwood PTR Prentice Hall
- R.G. Brereton, "Chemometrics - Data Analysis for the Laboratory and Chemical Plant", WILEY
- P. Gemperline (ed.) "Practical Guide to Chemometrics" CRC Press
- T. Naes, T. Isaksson, T. Fearn, T. Davies "A user-friendly guide to Multivariate Calibration and Classification" NIR Publications