Teaching code

4S004553

Teacher

Pietro Sala

Coordinator

Pietro Sala

Credits

6

Also offered in courses:

• Biomedical decision support systems of the course Master's degree in Artificial intelligence

Language

English

Scientific Disciplinary Sector (SSD)

INF/01 - INFORMATICS

Period

Semester 2 dal Mar 3, 2025 al Jun 13, 2025.

Courses Single

Authorized

Lessons timetable Moodle Seminars 0

Learning objectives

Knowledge and understanding The course aims to introduce principles that form the foundations of the Decision Support System together with case-studies of their real-world applications, with particular focus on their use in Biomedical domain. In particular, the purpose of the course consists of providing advanced knowledge on the techniques and principles involved in managing and manipulating very large databases (with specific examples borrowed from the biomedical domain). Moreover, the course will provide the theretical and practical foundations of the main data mining techniques used in clinical domains. Applying knowledge and understanding During the course students will aquire the following competences: - they will be able to choose and use the appropriate components in order to provide solution for supporting decision to the medical staff; - they will be able torealize complex operations of Extraction, Transformation, and Loading (ETL) on several clinical data types coming from different sources (Relational Databases, API, Websites, and so on) and encoded in both structure (relational tables) and semi-structured (XML) fashion; -they will be able to model and realize OLAP (On-Line Analytical Processing) solutions for supportuing decisions in a Biomedical context; -they will be able to use or adapt advanced data-mining techniques (Approximate Functional Dependencies, Association Rules, Entropy-based Classifiers, and so on) for extracting knowledge from large amounts of data. Making judgements Students will develop the required skills in order to be autonomous in the following tasks: - choose and apply data mining techniques for extracting medical knowledge from large amount of data; - choose the appropriate graphical/interactive representations for represent specific clinical information. Communication skills The student will learn how to address the correct priorities to the informations that must be reported to the end-user according to his needs and the language of his domain. Learning skills The students will be introduced to the main algorithms and techniques used in the clinical data mining field, together with the description of the factors that affect their efficiency and effectiveness. This knowledge will be the basis for comprehend more specific techniques adopted nowadays for data mining for clinical domain. Moreover, the student will be able to choose autonomously the data mining techinque for answering a given quesry of the end-user. Finally, he will be able to evaluate the performance and the accuracy of the proposed solution.

Prerequisites and basic notions

Possess solid knowledge of programming, including understanding of data structures and basic algorithms. Students should be familiar with fundamental concepts of statistics and probability, linear algebra, and mathematical analysis. Basic understanding of machine learning is required, including classification, regression, and clustering algorithms. Preliminary knowledge of databases and data management is essential. It is also important to have familiarity with graph theory concepts and algorithms, as they will be applied in pattern extraction and analysis of complex data structures.

Program

By the end of the course, students will be able to apply advanced data mining techniques for knowledge extraction from complex and heterogeneous datasets. They will master Association Rules algorithms to identify frequent patterns and significant relationships in data, using metrics such as support, confidence, and lift. They will develop skills in analyzing complex data structures through process mining and automata learning, implementing algorithms like L* for automaton learning. Students will acquire expertise in advanced machine learning techniques, including ensemble methods, boosting, conformal prediction, and univariate time-series analysis. Students will learn to apply game theory concepts in data mining, using Nash equilibria and Shapley values for model analysis and interpretation. Finally, they will develop skills in pattern mining, using entropy, feature selection, and conformal regression techniques to ensure robustness and reliability of predictions.

Bibliography

Didactic methods

The course adopts an approach primarily based on frontal lessons integrated with laboratory sessions using Jupyter Notebooks. Each module combines theoretical explanations with immediate practical implementations, allowing students to experiment directly with code and visualize results in real-time. Students will work with established industry tools and implement algorithms both from scratch and using specialized libraries. Educational resources include supplementary materials, example code, and access to libraries and frameworks in the field of artificial intelligence and data mining.

Learning assessment procedures

Learning assessment is based on assigned exercises during the course and a final oral examination. Students must complete all exercises proposed by the instructor, covering the different program topics: Association Rules, Mining Structures, Machine Learning, Game Theory for Data Mining, and Pattern Mining. No alternative proposals are provided - all assigned exercises are mandatory and must be submitted. During the oral examination, students will present the results of completed exercises and answer questions about the methods used, obtained results, and theoretical concepts underlying the applied data mining techniques.

Evaluation criteria

The evaluation is based on the quality of assigned exercise completion and the ability to orally discuss obtained results. Regarding exercises, the evaluation considers correctness of data mining technique implementation, accuracy in data analysis, and completeness of proposed solutions. During the oral examination, theoretical understanding of used algorithms, ability to critically interpret results, skill in connecting theory and practical application, and mastery of data mining technical language are assessed. Particular attention is given to the ability to justify methodological choices adopted in exercises and to discuss implications of obtained results.

Criteria for the composition of the final grade

The final grade is composed of 60% evaluation of assigned exercises and 40% oral examination. Exercise evaluation considers technical correctness of implementations, effectiveness of proposed solutions, completeness of result analysis, and presentation quality. Oral evaluation is based on presentation clarity, depth of theoretical understanding, and ability to justify implementation choices and conduct critical analysis of the project.

Exam language

Inglese