A. Mancuso, A. Iodice D'Enza, F. Palumbo, R. Simone

Telematics data allow large-scale assessment of how driving behaviour affects vehicle emissions, yet most estimation pipelines rely on complex machine learning models with limited interpretability. This study introduces an interpretable framework that combines emission reconstruction, surrogate modelling, and model-agnostic explanation techniques. SHapley Additive exPlanations are employed to break down trip-level emission estimates into individual feature contributions, while global summaries highlight the most influential behavioural patterns. In addition, Archetypal Analysis on SHAP values uncovers distinct emission-related driving profiles. This behavioural segmentation supports the design of targeted and scalable eco-driving strategies, enhancing both understanding and practical applicability of emission analysis. The proposed approach improves transparency and facilitates data-driven decision-making for sustainable mobility policies.

Keywords: Machine Learning, SHAP, Archetypal Analysis, Eco-driving, Telematics data

Scheduled

Machine Learning and Statistical Methods
September 2, 2026  3:30 PM
Aula 20


Other papers in the same session


Cookie policy

We use cookies in order to be able to identify and authenticate you on the website. They are necessary for the correct functioning of it, and therefore they can not be disabled. If you continue browsing the website, you are agreeing with their acceptance, as well as our Privacy Policy.

Additionally, we use Google Analytics in order to analyze the website traffic. They also use cookies and you can accept or refuse them with the buttons below.

You can read more details about our Cookie Policy and our Privacy Policy.