An international research team, including Emanuele Borgonovo, Director of the Department of Decision Sciences at Bocconi, Cynthia Rudin, the Gilbert, Louis, and Edward Lehrman Distinguished Professor of Computer Science at Duke University, Xuefei Lu assistant professor at Skema Business School (France), Francesco Cappelli at the University of Tuscia, and Elmar Plischke of the University of Clausthal (Germany), has been named one of the four finalists in the prestigious Best Paper competition of the Data Mining Society of INFORMS.

Their work introduces a novel method for determining the importance of variables in machine learning models while avoiding prediction errors due to extrapolation. This is an essential task to improve the explainability and interpretability of machine learning models.

The final took place at the Annual INFORMS Meeting in Seattle, where the team secured the Second Prize, surpassing teams from Caltech-Cornell and Tsinghua University, and finishing as runner-up to the University of Michigan team.

Bocconi's representation in the team is significant: Dr. Xuefei Lu and Dr. Francesco Cappelli graduated from Bocconi's PhD in Statistics program under the guidance of Prof. Emanuele Borgonovo.