Clinical practice has been experiencing a shift from population-based approaches to individualized strategies tailored according to the patient’s specific characteristics. A biomarker is a characteristic or variable that is objectively measured and that reflects the individual’s health status and risk at key time points. Nowadays, several biomarkers can be used to characterize the specific patient, obtained using different technologies, devices, and procedures. However, a complete overview of the patient’s status is often possible only looking at multiple dimensions simultaneously. Clinical indicators are summary measures that capture relevant information on different characteristics of the patient’s health status. They can be for different purposes such as identification of the risk of developing a disease, staging or classification of disease severity, prediction of future disease course, including recurrence, response to therapy, and monitoring efficacy of therapy. This course presents the main steps for the construction of clinical indicators, focusing on the most important Machine Learning tools suitable for this purpose. The course is divided in lessons, in which the theoretical concepts will be presented, and laboratory work, in which students will be asked to apply them to a real clinical problem.

None

1. Introduction and Knowledge Discovery from Databases 2. Data Cleaning and Data Reduction 3. Fuzzy Inference Systems 5. Artificial Neural Networks 6. Applications During the laboratory lessons the students will work in team and apply theoretical notions to the problem of weaning from mechanical ventilation.

A distanza in modalità sincrona

Sviluppo di project work in team

P.D.2-2 - Settembre

the course will take place in sept-oct 2024