基于自然语言处理的文本自动校对系统及实验结果分析

Automatic Text Proofreading System based on Natural Language Processing and Analysis of Experimental Results

设计了一种可以用于字词查错、语法查错、语义查错的文本自动校对系统。通过构建语料库,利用二元接续关系实现字词查错;基于规则库和语法成分算法,完成语法查错;根据依存关系和句子相似度计算,实现了语义查错。从实验结果来看,本研究设计的文本自动校对系统对错误句子的查准率在70%以上,能够快速、准确地找出文本中存在的错误,具有良好的应用前景。

Designs a text automatic proofreading system which can be used for word error checking,grammar error checking and semantic error checking.By constructing the corpus,the error checking of word is realized by using the binary continuity relation.Complete grammar error checking based on rule base and syntax component algorithm;Based on dependency and sentence similarity calculation,semantic error checking is realized.It can be seen from the experimental results that the accuracy rate of the automatic text proofreading system designed in this paper is more than 70%,which can find out the errors in the text quickly and accurately,and has a good application prospect.