Evaluation of Difficulty Estimation for Learning Materials Recommendation
Yasuo MIYOSHI, Kazuhiro Suzuki, Ken-ichi Shiota, Ryo OKAMOTO · International Conference on Computers in Education · 2014
The popular technology for the information recommendation of books or web pages is based on taste information from many users, but it is important to be based on difficulty and proficiency for the recommendation of learning materials. We have developed an algorithm to estimate difficulty of learning materials and proficiency of learners, for recommendation considering difficulty of learning materials. The algorithm uses only a bipartite learner-material graph that consists of the reader relations with materials and learners. In this paper, we describe how to make an accurate data to evaluate the estimated difficulty, and report about the result that evaluated the precision of our proposed algorithm.