期刊
IEEE ACCESS
卷 10, 期 -, 页码 118167-118175出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/ACCESS.2022.3219448
关键词
Korean grammar correction; error standard; gold test set; human evaluation
资金
- Ministry of Science and ICT, South Korea, under the Information Technology Research Center Support Program [ITP-2018-0-01405]
- Basic Science Research Program through the National Research Foundation of Korea, Ministry of Education [NRF-2022R1A2C1007616]
Recently, there has been active research on Korean grammatical error correction in machine translation and automatic noise generation. However, there is a lack of a gold-standard test set for objective and official comparative analysis. This study proposes a gold-standard test set called K-NCT for Korean grammatical error correction, which uses a new error type classification guideline to improve accuracy. Quantitative analysis using a commercialization system and human evaluation shows that the proposed test set is well-balanced, diverse, and precise.
Recently, active research has been conducted on Korean grammatical error correction on machine translation (MT) and automatic noise generation. However, there is no gold-standard test set for objective and official comparative analysis. A significant limitation is measuring the ill-defined performance because the experimental error types in the train set are also included in the test set. Moreover, error types in the training set are also included in the test set. Additionally, the types of errors for qualitative analysis are defined differently with no explicit guidelines. This study proposes a gold-standard test set called the Korean Neural Grammatical Correction Test set (K-NCT) for Korean grammatical error correction using a new error type classification guideline. To ensure the factuality and reliability of the proposal, we conduct a quantitative analysis using a commercialization system and human evaluation. Experimental results demonstrate that the proposed grammatical error correction test set has a well-balanced, diverse, and precise guideline. Our dataset is available at https://github.com/seonminkoo/K-NCT
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据