Title: Analysing behavioural and academic attributes of students using educational data mining

Authors: Muhammad Umer; Saima Sadiq; Arif Mehmood; Imran Ashraf; Gyu Sang Choi; Sadia Din

Addresses: Department of Computer Science, Khwaja Fareed University of Engineering and Information Technology, Rahim Yar Khan, 64200, Pakistan ' Department of Computer Science, Khwaja Fareed University of Engineering and Information Technology, Rahim Yar Khan, 64200, Pakistan ' Department of Computer Science & Information Technology, The Islamia University of Bahawalpur, Bahawalpur 63100, Pakistan ' Department of Information and Communication Engineering, Yeungnam University, Gyeongsan 38541, South Korea ' Department of Information and Communication Engineering, Yeungnam University, Gyeongsan 38541, South Korea ' Department of Information and Communication Engineering, Yeungnam University, Gyeongsan 38541, South Korea

Abstract: Educational data mining has attracted significant consideration over the last few years. Information that is stored online using educational systems is increasing tremendously. The online learning environment can be improved by analysing and mining this information to extract representative features about students' behaviour and academic skills. Various classifiers are investigated to analyse the prediction of students' academic performance based on the attributes from the Kalboard360 learning management system (LMS). The selection of significant features can substantially improve the prediction, hence, ANOVA and Chi-square filter approaches and forward feature selection and backward feature elimination wrapper approaches are examined for their efficacy. Results reveal that an extra tree classifier can achieve an accuracy of 0.8755 when trained on backward feature elimination selected features. Wrapper approaches prove to be effective in determining the most significant attributes for student performance prediction. 'Visited resources', 'raised hands', 'relation', 'parent answering survey', and 'student absent days' are regarded as the most significant attributes to determine student performance. Education specialists and institutions can leverage these findings to improve the student learning process and enhance their academic performance.

Keywords: metadata; educational data mining; feature extraction; ANOVA; chi-square; forward feature selection; backward feature elimination; extra tree classifier.

DOI: 10.1504/IJNT.2023.134005

International Journal of Nanotechnology, 2023 Vol.20 No.5/6/7/8/9/10, pp.451 - 476

Received: 18 Feb 2021
Accepted: 08 Apr 2021
Published online: 10 Oct 2023 *

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