Identifying dominant and heterogeneous factors influencing students’ future capabilities under the digital transformation of higher education: a machine learning approach

In the context of digital transformation of higher education, an in-depth analysis of the key factors influencing students’ future capabilities is crucial for optimizing curriculum design and resource allocation. This study combines Self-Determination Theory (SDT) with interpretable machine learning methods, and uses 1,083 valid questionnaires from the German Forum for Digitalization of Higher Education (HFD) from November 2023 to March 2024 as samples to construct and prune the J48 decision tree model to systematically identify the dominant and heterogeneous conditions of students’ future capabilities. The results show that the high-frequency use of diversified learning forms is the core driving force for improving students’ future capabilities. When the sense of autonomy is insufficient, high-quality learning space and positive evaluation of AI tools can effectively compensate for students’ intrinsic motivation. This study provides empirical evidence for educational practitioners and policymakers to cultivate students’ adaptability, innovation and interdisciplinary capabilities in digital learning environments.