Students Groups Detection in Online Examinations Using K-Means Clustering


  • Ahmad Al Musawi Department of Information Technology, University of Thi-Qar, Thi Qar,
  • Mustafa Asaad Hasan University of Thi-Qar
  • Ali Hussein Lazem University of Thi-Qar


K-Means Clustering, Online Examination, Students Group


Schools and universities have been adversely affected by the widespread distribution of COVID-19 and related quarantine around the worlds. As a result of this distribution, most of these institutions have deployed online teaching platforms as an alternative to students' physical attendance. However, the usages of recent online technologies have provided extra communication channels in addition to e-learning media. Data availability and accessibility have made it possible to conduct online searches. Studying the students' performance in the online examination is conducted to determine the degree of similarity and groups of students who shared similar behavior. The K-Means clustering model has been implemented on the tf-idf representation of the retrieved online corpus. The study concludes that students fall into five distinguished groups (i.e. small communities) based on similarity in performance of sharing the same significant content over the different courses. A larger corpus (document collection) of the complete academic performance of students at different levels (as future work) would help refine more accurate groups of collaboration among students.