抄録
本論文では,プログラミング教育において,学習者ひとりひとりの理解状況に応じて演習課題を出題する手法を提案し,それを実際のプログラミング演習科目へ適用した結果を示す.本手法では,豊富に用意された演習課題の中から,過去の他の学習者の演習履歴に基づいて各演習課題の達成度を推測し,各学習者に最適な演習課題を選出して出題する.学習意欲が低い学習者には学習の継続を,学習意欲が高い学習者にはプログラミング能力の向上を優先して出題を行うので,本手法は両者に対して効果的である.本手法を大学でのC言語プログラミング演習科目に適用した結果,選出された演習課題を解いた学習者の学習継続率が11.3ポイント向上し,理解をより深めさせることもできた.
This paper proposes a method to recommend a C-language programming exercise to each student according to its understanding. The understanding of each student for a specific programming exercise is assumed to be evaluated by a score. For a specific student, a score of a specific programming exercise is inferred, using score histories of himself and other students who addressed the same programming exercise sets in the past. The method selects an appropriate programming exercise to each student, based on the inferred score and the attitude toward the programming. It recommends an easy one to a student loosing motivation, while a tough one to a positive student. The application of the method to an actual university course has proved that students addressing recommended programming exercises are superior to ones rejecting the recommendation in the exercise continuity by 11.3 points, as well as the understanding.
