Educational Data Mining to support the Design of AI-Powered Tutoring Systems

Abstract

This paper explores the crucial role of data mining in enhancing AI-powered tutoring systems and hence educational systems in general. Data mining techniques enable the analysis of vast amounts of student data to uncover patterns, predict learning performance, and provide personalized educational experiences. By leveraging methods such as clustering, classification, regression, AI tutors can adapt to individual learning styles, recommend targeted resources and identify potential areas of difficulty and improvement. This paper suggests the integration of artificial intelligence and data mining in the field of education to improve learning efficiency, engagement, and academic success. It includes an empirical case study on data of 790 students, where data mining techniques were employed to reveal hidden patterns. The implementation was done using Orange Data Mining and Google Colab. The goal of this analysis is to support the development of AI-powered tutoring systems. This study represents a foundational step towards creating adaptive, data-driven tutoring models informed by a deeper understanding of student behavior. The paper compares the findings with previous research by one of the authors on different data to get more insights. It also makes recommendations to be implemented in AI tutoring systems based on the data mining results to enhance students’ motivation and to reduce the risk of drop out.