Developing E-Book Page Ranking Model for Pre-Class Reading Recommendation
Christopher C. Y. Yang, Gökhan AKÇ APINAR, Brendan Flanagan, Hiroaki Ogata · International Conference on Computers in Education · 2019
In this paper, we propose an E-Book Page Ranking (EBPR) method to rank e-book pages from the original learning material automatically. The proposed method ranks all the e-book pages by the class probabilities retrieved from machine learning models. The top-ranked e-book pages are then selected to form the pre-class reading (preview) recommendation. The proposed method extracts image features and text features from e-book page contents as well as the e-book usage features from students’ previous reading logs. In this paper, we test the performance of the proposed model with two different cases, with and without past e-book usage data. The experimental results showed the improvability of the model after taking into account learners’ past e-book usages.