基于描述逻辑本体推理的语义级中文校对方法

摘要

In recent years, the Chinese proofreading technology has achieved good results at the word-level and the grammar-level, while it's relatively weak at the semantic level. This paper provides a new method of Chinese proofreading powered by the ontology technology. The semantic contents are firstly extracted from the Chinese texts and transformed into some kind of structured ontology, which is combined with the correct background ontology. The logical consistency of the extracted semantic content is determined by description logic reasoning machine, with the detected logical consistency error mapped into some Chinese semantic errors. This method is tested in the domain of political sensitive information. Compared with other methods, it has obvious advantages of the transparency of the semantic proofreading, the integrity of the semantic model and the intelligence of the semantic reasoning.%近年来中文校对技术虽然在字词级和语法级层面取得了较好效果,但是对于语义级层面研究相对薄弱,目前相关研究都有一定局限性.提出利用本体技术将中文文本中的语义内容提取出来转换为结构化本体,再与正确的领域背景本体库融合,通过描述逻辑推理机来判断提取的语义内容的逻辑一致性,并将检测出的逻辑一致性错误映射为中文语义错误.此方法在政治敏感错误等领域进行了测试,与其他相关研究相比,具有语义查错的透明性、语义模型的完整性、语义推理的智能型等优势.