Student Performance Prediction Using AI and ML: State of the Art

Abstract

The digitalization of educational processes has enabled the generation of large datasets that can be used to improve processes in academic environments. One particular problem is the prediction of student performances based on historical data. Efficient student performance prediction can be used not only to prevent dropouts at an early stage, but it can also help perspective students to determine the fields in which they can have high academic performance and build successful student profile. Due to large and diverse data, this process has to be conducted with high degree of automations. Therefore, in this paper we have conducted an extensive survey on the impact of AI and ML techniques in student performance prediction, with primary aim to detect opportunities, good practices, but most importantly to identify gaps and remaining research challenges with the ultimate goal to define an effective framework for a student performance prediction system.