Comment
Author: Admin | 2025-04-28
Caused frustration among educators not knowing whether or how students were learning and engaging. This situation provided an unprecedented opportunity for PLA to show how AI and ML can help in gaining insights into student progress, needs, potential, and risk during a crisis or major changes. In the analyzed literature, the studies of (Abdullah et al., 2021; Dias et al., 2020) were found to specifically mention or address the role of PLA in the COVID-19 context. For instance, Dias et al. (2020) assessed students online learning interactions to predict the quality of their engagement during the pandemic. This modest extent of literature illustrating the response of PLA to the pandemic is expected to expand. It is likely that further findings and more detailed studies of the changes in PLA practice in the post pandemic era will emerge over time. Technically, the potential of PLA to manage situations caused by the pandemic is hindered by the so-called data drift (Adadi et al., 2021). In order to deliver prediction, models analyze historical data. However, pre-COVID data are no longer relevant since they do not include new student and teacher behaviors and interactions. For example, most of the predictions of the 2020–2021 admissions cycle were probably inaccurate, since the models based their prediction on historical data that does not reflect the impact of a global pandemic on student admissions and retention. ML models degrade gradually when the data they were trained on no longer reflect the present state of the world (Adadi et
Add Comment