Smart Signal: An Early Warning System to Prevent Student Drop-Out
Abstract
In Flanders, one in seven students leaves secondary school without a diploma, putting them at risk of long-term social and economic disadvantages. To help prevent such dropout, we develop an early warning system: Smart Signal. Drawing on the rich marketing literature on customer retention and on recent developments in machine learning, Smart Signal uses a hidden Markov model to capture longitudinal dynamics in student data from the school’s digital learning platform. Crucially, Smart Signal guards against algorithmic bias and stereotyping – common concerns when applying machine learning algorithms in education – thereby supporting fair and equitable outcomes. We propose a teacher-facing interface that frames Smart Signal as a co-pilot, not a replacement. We further evaluate teacher acceptance using an importance-performance map analysis. Our work highlights how marketing science and machine learning can contribute to socially impactful innovations in education while responding to real institutional constraints such as teacher burnout and resource scarcity.
About Tim Cosemans
Tim Cosemans is a data scientist and engineer at Smartschool, where he leads the development of Smart Signal, an early warning system that uses explainable AI to help prevent student drop-out. His work bridges technical innovation and educational impact, ensuring AI solutions are not only accurate but also ethical, interpretable, and aligned with school values.
Tim is also an affiliate researcher at Eindhoven University of Technology, where his research focuses on the intersection of data science, marketing analytics, and human behavior. He holds degrees in business engineering and statistical data analysis.
Across academia and industry, Tim is driven by a passion for making data meaningful, whether by guiding schools through digital transformation, supporting students with data-informed insights, or contributing to the broader conversation about responsible AI in education.
Language
English
This is a free seminar. Registration is mandatory.