TY - GEN
T1 - Predicting Student Dropout in Massive Open Online Courses Using Deep Learning Models - A Systematic Review
AU - Mbunge, Elliot
AU - Batani, John
AU - Mafumbate, Racheal
AU - Gurajena, Caroline
AU - Fashoto, Stephen
AU - Rugube, Talent
AU - Akinnuwesi, Boluwaji
AU - Metfula, Andile
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Predicting student dropout is becoming imperative in online learning platforms. Before COVID-19, predicting student dropout was systematically done manually. Therefore, there is a need to analyse students’ behaviour, cognitive learning styles, and other metacognitive patterns of learning in real-time from available data repositories to reduce student dropout and subsequently develop robust strategies for instructional design and remedial interventions to enhance student success and retention. In this study, we present a comprehensive review of deep learning models applied to predict student dropout in online learning platforms. In addition, challenges and opportunities associated with online learning are presented in this study. The study revealed that convolutional neural networks, recurrent neural networks, long short-term memory and bidirectional long short-term memory have been predominantly used to predict student dropout using predictors such as course assessments, socio-economic, access to online resources, personal skills and course attributes. However, the study revealed that the psychological state of students was not taken into consideration by many authors, yet it impacts students’ learning outcomes and assists policymakers in providing remedial interventions. Therefore, future work can delve deeper into the integration of psychological attributes such as stress, anxiety, attitude towards studying, student interests and counselling sessions to predict student dropout during disasters and health emergencies.
AB - Predicting student dropout is becoming imperative in online learning platforms. Before COVID-19, predicting student dropout was systematically done manually. Therefore, there is a need to analyse students’ behaviour, cognitive learning styles, and other metacognitive patterns of learning in real-time from available data repositories to reduce student dropout and subsequently develop robust strategies for instructional design and remedial interventions to enhance student success and retention. In this study, we present a comprehensive review of deep learning models applied to predict student dropout in online learning platforms. In addition, challenges and opportunities associated with online learning are presented in this study. The study revealed that convolutional neural networks, recurrent neural networks, long short-term memory and bidirectional long short-term memory have been predominantly used to predict student dropout using predictors such as course assessments, socio-economic, access to online resources, personal skills and course attributes. However, the study revealed that the psychological state of students was not taken into consideration by many authors, yet it impacts students’ learning outcomes and assists policymakers in providing remedial interventions. Therefore, future work can delve deeper into the integration of psychological attributes such as stress, anxiety, attitude towards studying, student interests and counselling sessions to predict student dropout during disasters and health emergencies.
KW - Deep learning
KW - Massive open online courses
KW - Prediction
KW - Student dropout
UR - http://www.scopus.com/inward/record.url?scp=85135063813&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-09073-8_20
DO - 10.1007/978-3-031-09073-8_20
M3 - Conference contribution
AN - SCOPUS:85135063813
SN - 9783031090721
T3 - Lecture Notes in Networks and Systems
SP - 212
EP - 231
BT - Cybernetics Perspectives in Systems - Proceedings of 11th Computer Science On-line Conference, CSOC 2022, Vol 3
A2 - Silhavy, Radek
PB - Springer Science and Business Media Deutschland GmbH
T2 - 11th Computer Science On-line Conference, CSOC 2022
Y2 - 26 April 2022 through 26 April 2022
ER -