難易度特定要素に基づいた英語短文穴埋め問題提示システム

摘要

英語の短文穴埋め問題を解くためには文法や語彙など複数の知識が必要である．学習者によって得意・不得意とする知識が異なるため,大量の問題の中から学習者の状態に適した問題を選ぶことは困難である．本研究では，難易度に影響を与える要素灘易度特定要素）に基づいて段階的に知識を習得できる短文穴埋め問題提示手法を提案する.段階的に問題間の関係を表現するため，個々の難易度に沿って問題を整理した間席ネットワークを構築する．学習者の学習結果に応じて問題ネットワークを辿ることで難易度特定要素を段階的に学習できる．実験結果から学習者の状態に応じた問題が提示できていることを確認した．%Learning English with multiple-choice cloze questions is effective for checking knowledge of vocabulary and grammar. Since important knowledge and questions which should be learned are different depending on learners, selecting adequate questions from a large amount of questions is too difficult. It is necessary to support learners gradually by showing the difficult questions according to the understanding level of the learner. In this research, we specify features which affect to the difficulty of multiple-choice cloze questions and propose a method for presenting questions based on these features. In our method, a question network where nodes are question sets with the same difficulty related features and links represent sequential difficulties is constructed. Questions are selected in question network based on learner's understanding level. From the experimental result, we confirm our method is effective for stepwise learning.