Systematic Literature Review: Machine Learning in Education to Predict Student Performance

Abstract

Machine learning can be beneficial to numerous fields, and of which is education. A particular area, which is predicting student performance, is selected in this paper to be reviewed. The objective of this paper is to find out what predictors can be used to predict student performance, what performances of students should be predicted, and what algorithms can be used for predicting student performance and which one has the best accuracy. To achieve those objectives, a systematic literature review has been done. In this research, it is found that there are various student features that can be used as predictors, student performance indicators that can be predicted with machine learning, and there are also a variety of machine learning algorithms that can be used in predicting student performance. Regarding the machine learning algorithm with the best accuracy/performance, the results differ from one research with the other. It is also found that increasing the number of predictors used and implementing techniques such as feature selection and data mining can improve accuracy of the prediction. We conclude that in predicting student performance, there are various predictors predicted student performance indicator, and machine learning algorithms that can be applied, as well as results regarding the most accurate machine learning algorithm differ.