基于依存句法分析的数学表达式查询扩展方法

MATHEMATICAL EXPRESSION QUERY EXPANSION METHOD BASED ON DEPENDENCY PARSING

  • 摘要: 传统的数学表达式检索方法主要面向表达式的二维结构, 难以检索出具有相同语义但结构不同的数学表达式。针对这一问题, 设计一种基于依存句法分析的数学表达式查询扩展方法。以FDS(Formula Description Structure)解析表达式结构信息的检索方法为基础, 通过依存句法抽取表达式周围文本中的语义词, 并建立运算符、表达式、语义词间索引; 通过语义词二次查询, 实现检索语义相同但结构不同数学表达式的目的。实验结果表明, 依存分析能有效地抽取数学表达式语义词, 加入语义的表达式检索方法, 查全率和查准率都有了一定的提高。

     

    Abstract: Traditional methods of mathematical expression retrieval mainly focus on the two-dimensional structure of expressions. It is difficult to retrieve mathematical expressions with the same semantics but different structures. Aimed at this problem, a mathematical expression query expansion method based on dependency parsing is designed. Based on the retrieval method of formula description structure (FDS) analytical expression structure information, the semantic words in the text surrounding the expression were extracted through the dependency syntax, and the index between operators, expressions, and semantic words was established. Through the second query of the semantic words, the purpose of retrieving mathematical expressions with the same semantics but different structures was realized. The experimental results show that dependency analysis can effectively extract the semantic words of mathematical expressions, and the semantic expression retrieval method is added at the same time, and the recall rate and precision rate are improved to a certain extent.

     

/

返回文章
返回