Study on information presentation method for misuse detection and correction in Japanese writing support system

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K01114
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Educational technology
• Research Institution: National Institute of Informatics
• Principal Investigator: Abekawa Takeshi 国立情報学研究所, コンテンツ科学研究系, 特任研究員 (00431776)
• Co-Investigator(Kenkyū-buntansha): ホドシチェク ボル 大阪大学, 言語文化研究科(言語文化専攻), 講師 (10748768)
• Co-Investigator(Renkei-kenkyūsha): NISHINA Kikuko 東京工業大学, 名誉教授 (40198479)
• Research Collaborator: UTASHIRO Takafumi ; YAGI Yutaka ; SOU Kosen ; Irena Srdanovic
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 日本語学習支援 / 作文支援 / 接続表現 / 誤用検出 / アカデミックライティング

Outline of Final Research Achievements

In this research project, we focused on "connection expression" from various grammatical elements which Japanese learner is liable to misuse, and constructed a "connection expression dictionary". As a concrete method, all expressions within 5 morphemes are collected from the beginning of sentence to the reading point in large-scale balanced corpus and science and technology paper collection. And we manually assigned labels of whether or not to express connection to frequently used expressions. As a result, we acquired 568 expressions as connection expressions. We conducted various analysis on the obtained connection expressions. and we showed that it can be obtained mechanically for medium-frequency connection expressions. In the future, we would like to incorporate the obtained connection expressions in the composition support system, to detect misuse and present candidates for paraphrasing.

Free Research Field

自然言語処理