DATA MINING ALGORITHMS FOR PREDICTION OF STUDENT TEACHERS’ PERFORMANCE IN ICT: A SYSTEMATIC LITERATURE REVIEW

Abstract

Poor ICT performance in teacher training colleges makes it more difficult for the majority of teachers to successfully use ICT resources in their teaching and learning. When teachers can efficiently utilize ICT resources, it empowers them to update their knowledge through online learning, consequently enhancing the overall quality of teaching and learning. This positive outcome can be observed through improved ICT performance. The aim of this article is to identify the appropriate Data Mining algorithms for predicting student teachers’ performance in ICT. The systematic literature review that was guided by the PRISMA statement 2020 served as the study methodology. It makes for clear reporting and offers a detailed checklist and flow diagram that direct the review procedure. On November 6, 2022, about 196 scholarly articles were downloaded from three digital libraries: Science Direct (38), ACM Digital Library (72), IEEE Xplore (51), and 35 from the Google Scholar search engine. After screening and eligibility checking, 28 scholarly articles were selected and analysed through content analysis in terms of the most commonly used algorithms, the year of publication, the study purposes, and the accuracy performance metrics. Considering the specific study findings represented quantitatively, Decision Trees and Naive Bayes were found to be the most commonly used Data Mining algorithms, with a count of 20.6% each. The most recently identified articles were published between 2014 and 2022. In terms of study purposes, a large number of studies focused on predicting student performance. Furthermore, about 6 out of 8 algorithms used in previous studies were found to score 80% or above in the average percentage of the highest and lowest accuracy metrics. Therefore, considering the general findings, the study identified five Data Mining algorithms as appropriate and most commonly used for prediction of student teachers’ performance in ICT. They are Naive Bayes, K-Nearest Neighbour, Support Vector Machine, Random Forest, and Decision Tree. The findings of this study would assist the government, college tutors, and student teachers in making better decisions to improve ICT performance for pre-service and in-service teachers.