TY - GEN
T1 - A Deep Learning Model for MOOC Dropout Prediction Using Learner's Course-relevant Activities
AU - Sultan, Mohamad T.
AU - Sayed, Hesham El
AU - Khan, Manzoor Ahmed
AU - Abduljabar, Mohammed
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Today, Open Massive Online Courses (MOOCs) have become very popular learning platforms with millions of participants. MOOCs provide a flexible distance-learning style courses usually delivered by top international universities. However, despite all benefits and features of MOOCs, these platforms have been heavily criticized due to students' high dropout rate. This have become a phenomenon on MOOCs, where users may enroll in a course but most of these users will dropout the course somewhere before the end. This has triggered the need for a development of a reliable and efficient dropout prediction model that can address this problem and maintain an encouraging learning activity. In this research, we present a deep leaning dropout predictor model to address this classification problem. By observing learner's early course activities and extensive feature engineering, we tried to predict the likelihood of student MOOCs dropout by using deep learning artificial neural networks (ANNs). Through selecting the best parameter values and using validation approach, our model was able to achieve 91% in terms of precision and 90% in terms of accuracy which are better than existing studies. Our obtained results are compared and benchmarked against the existing state of the art literature that addresses the same problem.
AB - Today, Open Massive Online Courses (MOOCs) have become very popular learning platforms with millions of participants. MOOCs provide a flexible distance-learning style courses usually delivered by top international universities. However, despite all benefits and features of MOOCs, these platforms have been heavily criticized due to students' high dropout rate. This have become a phenomenon on MOOCs, where users may enroll in a course but most of these users will dropout the course somewhere before the end. This has triggered the need for a development of a reliable and efficient dropout prediction model that can address this problem and maintain an encouraging learning activity. In this research, we present a deep leaning dropout predictor model to address this classification problem. By observing learner's early course activities and extensive feature engineering, we tried to predict the likelihood of student MOOCs dropout by using deep learning artificial neural networks (ANNs). Through selecting the best parameter values and using validation approach, our model was able to achieve 91% in terms of precision and 90% in terms of accuracy which are better than existing studies. Our obtained results are compared and benchmarked against the existing state of the art literature that addresses the same problem.
KW - ANN
KW - Dropout prediction
KW - MOOCs
KW - deep learning
UR - http://www.scopus.com/inward/record.url?scp=85147661995&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85147661995&partnerID=8YFLogxK
U2 - 10.1109/GCAIoT57150.2022.10019063
DO - 10.1109/GCAIoT57150.2022.10019063
M3 - Conference contribution
AN - SCOPUS:85147661995
T3 - 2022 IEEE Global Conference on Artificial Intelligence and Internet of Things, GCAIoT 2022
SP - 13
EP - 18
BT - 2022 IEEE Global Conference on Artificial Intelligence and Internet of Things, GCAIoT 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2022 IEEE Global Conference on Artificial Intelligence and Internet of Things, GCAIoT 2022
Y2 - 18 December 2022 through 21 December 2022
ER -