@article{oai:ipsj.ixsq.nii.ac.jp:00220472, author = {佐々木, 健太 and 鈴木, 健一 and 君島, 朋子 and 吉川, 将司 and 乾, 健太郎 and Kenta, Sasaki and Kenichi, Suzuki and Tomoko, Kimijima and Masashi, Yoshikawa and Kentaro, Inui}, issue = {3}, journal = {情報処理学会論文誌教育とコンピュータ(TCE)}, month = {Oct}, note = {我々は大学・大学院のレポート採点に利用可能なレポート採点支援システムを開発した.具体的には,自然言語処理に基づく類似文検索の技術を利用し,採点対象となるレポート本文から採点基準を満たす可能性がある文を事前に抽出する.そして,抽出した文をレポート採点時にハイライト表示することで,教員のレポート採点作業を支援する.開発したシステムをグロービス経営大学院に導入したところ,従来の採点方法と比較して採点作業に掛かる時間を短縮しただけでなく,教員による採点結果のばらつきも低減した.さらに,採点にともなう教員の心理的負担を軽減できる可能性があることまで分かった.そして,採点基準に基づいて採点するレポートであれば,他の大学・大学院のレポート採点にも活用できる可能性がある., We have developed a report grading support system that can be used for grading written reports at universities and graduate schools. Specifically, the system extracts sentences that could satisfy the grading criteria from the text of the report in advance, using similar sentence retrieval techniques based on natural language processing. Then, the extracted sentences are highlighted in grading the report in order to assist faculty members. We introduced the developed system to the Graduate School of Management, Globis University. The experimental results showed that it not only shortened the required time for grading written reports, but also reduced the variability of grading results among faculty members, compared to the conventional grading method. Furthermore, we found that it has the potential to reduce the psychological burden that grading has on faculty members. In addition, if the reports are graded based on the grading criteria, there is a possibility that the system can be used for grading written reports in other universities and graduate schools.}, pages = {35--49}, title = {大学・大学院におけるレポート採点支援システムの開発}, volume = {8}, year = {2022} }