Using Explainable Machine Learning to Automatically Provide Feedback to Students Based on Data Analysis

Abstract

Providing feedback to students is one of the most powerful practices that have enhanced education in the world today. Despite there being useful feedback provided by students’ self-regulation and teachers’ feedback provision, there is still a need for feedback that provides meaningful insights or actionable information about the reasons behind it, which is not provided by the said feedback. This paper explores how we can use explainable machine learning to compute data-driven feedback concerning students’ academic performance and generate actionable recommendations which are beneficial for students and teachers. This method has been developed based on LMS (Learning Management System) data from a university course. The effectiveness of the proposed approach has been evaluated with the results demonstrating 90% accuracy.