阅读说明：本技术 一种基于反向翻译的英文文章自动语法纠错方法 (English article automatic grammar error correction method based on reverse translation ) 是由 杨东强 孙晓东 于 2021-07-26 设计创作，主要内容包括：本发明公开了一种基于反向翻译的英文文章自动语法纠错方法方法,通过分析现有学习者语料的错误类型分布,对常见的错误类型如动词、名词、部分介词、拼写和标点等建立混淆集。首先使用混淆集结合替换规则的方法对单语语料数据进行加噪处理,并将其用于训练基于反向翻译的自动错误生成模型。其次,通过将语法纠正模型产生的中间数据反馈输入到错误生成模型,并进行联合训练。最终通过优化的错误生成模型,输出接近学习者语料库的更高质量的合成训练语料,提高了纠错模型的泛化能力,帮助用户减少语法错误的发生率。(The invention discloses an English article automatic grammar error correction method based on reverse translation, which establishes a confusion set for common error types such as verbs, nouns, partial prepositions, spellings, punctuations and the like by analyzing the error type distribution of the corpus of the existing learners. Firstly, a method of combining an confusion set with a replacement rule is used for carrying out noise adding processing on the monolingual corpus data, and the monolingual corpus data is used for training an automatic error generation model based on reverse translation. And secondly, feeding back and inputting intermediate data generated by the grammar correction model to the error generation model, and performing joint training. Finally, through the optimized error generation model, the synthesized training corpus with higher quality close to the learner corpus is output, the generalization capability of the error correction model is improved, and the user is helped to reduce the incidence rate of grammar errors.)