Academic Performance Prediction Based on Multisource, Multifeature Behavioral Data

Digital data trails from disparate sources covering different aspects of student life are stored daily in most modern university campuses. However, it remains challenging to (i) combine these obtain a holistic view student, (ii) use accurately predict academic performance, and (iii) such predictions promote positive engagement with the university. To initially alleviate this problem, article, model named Augmented Education (AugmentED) is proposed. In our study, (1) first, an experiment conducted based on real-world campus dataset college students (N =156 ) that aggregates multisource behavioral not only online offline learning but also behaviors inside outside classroom. Specifically, gain in-depth insight into features leading excellent or poor metrics measuring linear nonlinear changes (e.g., regularity stability) lifestyles estimated; furthermore, representing dynamic temporal lifestyle patterns extracted by means long short-term memory (LSTM). (2) Second, machine learning-based classification algorithms developed performance. (3) Finally, visualized feedback enabling (especially at-risk students) potentially optimize their interactions achieve study-life balance designed. The experiments show AugmentED can students' performance high accuracy.