Using Knowledge Graph for Explainable Recommendation of External Content in Electronic Textbooks

Over the last 10 years, the world experienced a rapid increase in volume and diversity of digital learning resources. The abundance of digital resources could support a range of powerful educational scenarios, which were not available before. In this paper, we introduce a novel approach that combines fully automatic knowledge modeling, student modeling, and content recommendation approaches to recommend relevant Wikipedia articles for students working with online electronic textbooks. An assessment of our approach with real classroom data indicated several benefits of our approach over the baseline and revealed interesting patterns of students’ behavior while using the system.