摘要：

以往的中文文本查错研究主要针对字词错误，对句法、语义的错误推理研究不够。为此，利用统计模型和大规模人民日报语料库构建并扩充查错知识库，针对文本中字词、语法以及语义3个层次的错误，提出相应的多层级查错推理模型。设计并实现3个层级的文本查错算法，构建自动查错系统进行综合查错。实验结果表明，该系统查错性能较优，召回率达到85.62%。

关键词: 中文文本\知识库, 多层级查错, 查错推理, 语义查错

Abstract:

The previous search on text error-detecting focused on words-level, and the syntax and semantic errors reasoning are not paid enough attention. Words, syntax and semantic knowledge bases are enlarged and constructed by taking advantage of statistic model and massive People’s Daily corpus to resolve the relevant level error reasoning, and the relevant multi-level reasoning model is put forward. The comprehensive three-level text proofreading algorithm is implemented. The constructed system can check various text errors. Experimental results show that the system has a better performance, and the recall rate is 85.62%.

Key words: Chinese text, knowledge bases, multi-level error-detecting, error-detecting reasoning, semantic error-detecting

中图分类号: 

• TP317