Browsing by Author "Pretorius, Andre"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Towards a learning analytics reference framework to predict at-risk students at the Faculty of Military Science
    (Stellenbosch : Stellenbosch University, 2022-12) Pretorius, Andre; Khoza, Lindiwe Mhaka; Dalton, Wayne Owen; Stellenbosch University. Faculty of Military Science. School for Geospatial Studies and Information Systems.
    ENGLISH ABSTRACT: Learning analytics (LA) is a relatively new field of application in the Analytics domain. Its main aim is to analyse teaching and learning (T&L) data from various sources to provide users with insights towards improving T&L. One of these T&L improvements is a greater focus on student success and more accurate methods of limiting student failure. This process starts with the identification of students at risk of failure (so-called “at-risk” students) through a prediction methodology which commonly falls within the knowledge sphere of Artificial Intelligence (AI), more specifically Machine Learning (ML). In contemporary information systems, the supporting platform for this is provided by an LA information system (LAIS) that relies on an underlying virtual learning environment (VLE), which in turn uses T&L data from a learning management system (LMS). A reference framework (RF) establishes a common foundation for future implementation of a system for developers and users. It provides appropriate guidance to users in a specific field of knowledge. Guidance is, however, generic in nature to secure reusability. This research focussed on developing an RF to implement LA in the Faculty of Military Science (FMS) of Stellenbosch University (SU) for at-risk student identification. The RF is supported by five models and one framework, namely, (1) a pedagogical model, (2) a model for effective VLEs, (3) a model for LA implementation, (4) a model for at-risk student identification and (5) a framework for the ethical use of LA. It is the conclusion of the study that the RF for LA in the FMS will provide suitable guidance for future implementation of LA in the faculty to effect timely identification of at-risk students and fitting remedial actions towards greater throughput may be implemented. It is envisioned that this RF be validated in the FMS in the near future and that future research in the use of ML be extended to identify suitable indicators of at-risk students more accurately.