Construction and performance evaluation of an automatic scoring system for essay sentences by the intelligence mining

2018 Fiscal Year Final Research Report

• Project/Area Number: 16K00299
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Intelligent informatics
• Research Institution: Shizuoka University
• Principal Investigator: Hasegawa Takahiro 静岡大学, 情報基盤センター, 准教授 (40293609)
• Research Collaborator: HASEGAWA HIROAKI
• Project Period (FY): 2016-04-01 – 2019-03-31
• Keywords: マイニング / 文章力 / 形態素分析 / 学術認証フェデレーション / 情報基盤 / インテリジェンスマイニング

Outline of Final Research Achievements

Use morpheme analysis to decompose into 11 parts of speech such as nouns, verbs, particles, and auxiliary verbs included in short essays (about 800 to 2000 characters) We evaluated the short essay high-speed automatic scoring cloud service for scoring. According to multiple student questionnaires, the combined proportion of “Evaluation 5: Very effective” and “Evaluation 4: Effective” is 64% to 83% in the five-level evaluation of the effectiveness of the service, and many subjects used this service It turned out that the effect of the improvement of the writing ability by can be realized at a high rate. The effectiveness of the service in higher education has been verified with a high degree of certainty.

Free Research Field

情報システム

Academic Significance and Societal Importance of the Research Achievements

本研究によって日本語文章力を診断（採点）し、文章改善の要点を指摘する小論文高速自動採点クラウドサービスを、１万名を超える国立大学法人で利用可能にした。また、同システムを、学認機構を採用する国内外の数百の大学や研究機関において導入を容易にした。研究出力のために、大量の文章を出力する学術機関において、教員と学生間の文章添削の作業時間を大幅に削減できる効果を与えることが期待される。