TamilSpell: A Tamil spelling error corpus
收藏DataCite Commons2024-09-02 更新2025-04-09 收录
下载链接：
https://discovery.dundee.ac.uk/en/datasets/tamilspell-a-tamil-spelling-error-corpus
下载链接
链接失效反馈官方服务：
资源简介：
TamilSpell is a comprehensive four-column corpus that includes the sentence or word with errors, the corrected version of the sentence, the error code or tag, and the edit distance. The corpus comprises a total of 10,157,662 entries, encompassing 8,858,630 isolated errors falling under three categories (non-word errors, real-word errors and sandhi errors), 803,531 errors in 2-way combinations, and 495,501 errors in 3-way combinations.
1. Isolated errors (ISOLATED ERRORS.txt)
2. 2-way combinations (RWE+NWE.txt, RWE+SE.txt and SE+NWE.txt)
3. 3-way combinations (SE+RWE+NWE.txt)
提供机构：
University of Dundee
创建时间：
2024-08-30
