英语短文自动分析关键问题及方法的研究

When the English essays of Chinese students are analyzed by the methods of related researches, there are the problems that their correcting accuracy of grammar mistakes is low, their correcting types of grammar mistakes are few, as well as their credibility and validity are unsatisfactory in sentiment analysis and coherence analysis. Therefore, the research proposal will focus on three key problems that the correcting of grammar mistakes, analyzing of sentiment trend and discourse coherence in English essays of Chinese students. Finally, it will design some methods to analyze English essays of Chinese students, such as building the rules library of grammar mistakes for common grammar mistakes in writing of Chinese students, designing the correcting model of grammar mistakes combining rule and statistical methods for writing of Chinese students; building the context sentiment word library based on texts of Chinese students, researching the context vector representations and calculating methods of word and sentence based on texts of Chinese students, designing the analysis model of sentiment trend for English essays of Chinese students; building the chain of entity-based gird for entity word and its grammatical role based on texts of Chinese students, researching the methods of extracting coreference features and coreference resolution based on English texts, designing the analysis model of discourse coherence for English essays of Chinese students.

为了解决相关研究在分析中国学生英语短文时，语法错误纠正精度不高与纠错类型不多，情感倾向与语篇连贯分析信度与效度不理想的问题。本项目总结了中国学生写作易犯语法错误规则，构建了英语语法错误规则库，设计了规则与统计相结合的中国学生英语语法错误纠正模型；构建了中国学生英语文本上下文关系的情感词库，研究了英语短文上下文关系的单词句子情感向量表示与情感极性计算方法，设计了中国学生英语短文情感倾向分析模型；构建了中国学生英语文本实体词及其语法角色的实体网格链，研究了英语文本的共指特征提取与共指消解方法，设计了中国学生英语短文语篇连贯分析模型。

为了解决相关研究在分析中国学生英语短文时，语法错误纠正精度不高与纠错类型不多，情感倾向与语篇连贯分析信度与效度不理想的问题。本项目总结了中国学生写作易犯语法错误规则，构建了英语语法错误规则库，设计了规则与统计相结合的中国学生英语语法错误纠正模型；构建了中国学生英语文本上下文关系的情感词库，研究了英语短文上下文关系的单词句子情感向量表示与情感极性计算方法，设计了中国学生英语短文情感倾向分析模型；构建了中国学生英语文本实体词及其语法角色的实体网格链，研究了英语文本的共指特征提取与共指消解方法，设计了中国学生英语短文语篇连贯分析模型。