Title: An improvement of a data mining technique for early detection of at-risk learners in distance learning environments

Authors: Safia Bendjebar; Yacine Lafifi; Hassina Seridi-Bouchelaghem

Addresses: LabSTIC Laboratory, University 8 May 1945 Guelma, P.O. Box 401, 24000 Guelma, Algeria ' LabSTIC Laboratory, University 8 May 1945 Guelma, P.O. Box 401, 24000 Guelma, Algeria ' LabGED Laboratory, University Badji Mokhtar Annaba, P.O. Box 12, 23000 Annaba, Algeria

Abstract: The coronavirus pandemic has spread to several countries resulting in one of the largest educational disruptions in history. To address this challenge, the use of distance education in many institutions is implemented by adopting online learning platforms. The administrators of these platforms are scared off by the high number of at-risk learners. Early prediction of these learners can allow instructors to encourage them to complete their classes. Several works have explored data mining techniques to detect learners' failures. Our study is different from the existing ones in different ways: 1) it is a new approach based on the dynamic profiles of the learners; 2) it seeks to determine whether the use of a more accurate classification can be useful or not; 3) it provides a systematic comparison with other methods and works. The effectiveness of our prediction technique is assessed through the use of real data gathered from computer science courses.

Keywords: student prediction; student failure; educational data mining; EDM; classification; K nearest neighbours.

DOI: 10.1504/IJKL.2022.121958

International Journal of Knowledge and Learning, 2022 Vol.15 No.2, pp.185 - 202

Received: 18 Dec 2020
Accepted: 23 Aug 2021

Published online: 07 Apr 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article