Personalized Navigation Recommendation for E-book Page Jump

As the utilization of digital learning materials continues to rise in higher education, the accumulated operational log data provide a unique opportunity to analyze student reading behaviors. Previous works on reading behaviors for e-books have identified jump-back as frequent student behavior, which refers to students returning to previous pages to reflect on them during the reading. However, the lack of navigation in e-book systems makes finding the right page at once challenging. Students usually need to try several times to find the correct page, which indicates the strong demand for personalized navigation recommendations. This work aims to help the student alleviate this problem by recommending the right page for a jump-back. Specifically, we propose a model for personalized navigation recommendations based on neural networks. A two-phase experiment is conducted to evaluate the proposed model, and the experimental result on real-world datasets validates the feasibility and effectiveness of the proposed method.