基于OAVU四元组的数值型知识元表示方法与实践

Numerical Knowledge Element Representation Method and Application Based on OAVU Quadruple Structure

[目的/意义]以论文、专利为代表的科技文献数值信息是技术科研活动和科技进展的客观反映,充分挖掘这些数据信息的价值,发现其中的动态变化,对技术发展规律掌握与未来演化趋势预测具有重要的现实意义与研究价值。[方法/过程]针对数值型知识元知识表达方法相关研究的不足,本文创新性地设计出“研究对象—属性—数值—单位”(Object-Attribute-Value-Unit,OAVU)数值型知识元四元组表示结构,然后根据特定领域数值型知识元的词汇表达、句法特征与表达方式构造相对应的正则表达式实现数值型知识元的抽取。[结果/结论]以锂离子电池磷酸铁锂正极材料研究领域为分析对象,验证了本文所提出的数值型知识元四元组结构较知识工程三元组结构在知识抽取结果方面更加准确,弥补了电化学领域表征技术发展水平的数值型知识元知识表示方法的不足,为科技文献大数据的开发利用奠定了数据基础。

[Purpose/Significance]Numerical information represented by papers and patents was an objective reflection of technology research activities and progress of science and technology,and fully mining the value of numerical information and finding data dynamic change has important practical significance and research value for mastering technology development rule and forecasting future evolution trend.[Method/Process]In view of the shortcomings of relevant researches on numerical knowledge element knowledge expression methods,the paper innovatively designed the“Object-Attribute-Value-Unit”(OAVU)numerical knowledge element quadruple expression structure,and then a corresponding regular expression was constructed according to the lexical expression,syntactic characteristics and expression mode of the numerical knowledge element.[Result/Conclusion]Taking lithium iron phosphate cathode material for lithium ion battery as the analysis object,the numerical knowledge element quadruple representation structure was more accurate on knowledge extracting results than knowledge engineering triplet representation structure,making up for the deficiency of numerical knowledge element knowledge representation methods of reflecting technology development level,and providing a data foundation for the development and utilization of big data of scientific and technological literature.