After AY2020, the new "national center test for university admissions" will start. Description-type questions that are not included in the current national center test are newly introduced. The marking process for description type problems -- evaluating scores for each answer script -- is modeled as the semantic similarity of answer sentences to the correct answer sentences. It can be handled within a framework of natural language analysis. In this paper, based on natural language analysis, we proposed a basic model and a prototype method for automated marking. We also discussed some problems regarding the actual implementation of this method.
内容記述
原著
2017年度神奈川大学総合理学研究所共同研究助成論文
雑誌名
Science Journal of Kanagawa University
巻
29
ページ
39 - 45
発行年
2018-06-30
ISSN
1880-0483
書誌レコードID
AA12068302
著者版フラグ
publisher
その他の言語のタイトル
Proposal of Model for Automated Marking Techniques for Description Type of Examination