A data-analytical framework for the early detection of at-risk students in higher education

Data analytics has become one of the emerging tools for the future of education. By analyzing aggregated data from various sources, educators can identify at-risk students who are struggling in their courses and apply a variety of interventions to support their learning. However, despite the abundance of educational data available in universities, accurately identifying students at risk of poor performance in a course remains challenging. In a study conducted at a local university in Hong Kong, 90 students in a senior-year psychology course were involved. We employed a novel data-analytics approach that combined LASSO (Least Absolute Shrinkage and Selection Operator) regression and the Youden index to predict student performance and identify potentially at-risk students in the course. Additionally, we developed an open-source Python package (https://pypi.org/project/dualPredictor/) based on our method. This tool enables educators to easily apply advanced analytics techniques to their datasets, enhancing the accessibility of technology in education. This work underscores the transformative potential of data-driven, learner-centered approaches in higher education. Copyright © 2026 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.