Student performance prediction in higher education: A comprehensive review
Ellysa Tjandra, Sri Suning Kusumawardani, Ridi Ferdiana
Abstract
Student dropout still becomes a critical problem in education. Educational Data Mining (EDM) can bring potential impact to support academic institution’s goals in making academic decisions, such as regulation renewal, rule enforcement, or academic process improvement. The sooner at-risk students can be identified, the earlier institution members can provide necessary treatments, thus prevent them from dropout and increase the student retention rate. This study performs a comprehensive literature review of student performance prediction using EDM techniques, including various research from 2002 to 2021. Our study is aimed to provide a comprehensive review of recent studies based on student performance prediction tasks, predictor variables, methods, accuracy, and tools used in previous works of student performance prediction. Performing student performance prediction in an academic institution can be helpful to provide the student performance mitigation mechanism because it can be managed earlier by the management to decrease the student dropout rate.