利用SemRep语义网及MeSH语义网表达单篇论文知识

Detection of knowledge expression in a single paper using SemRep semantic network and MeSH semantic network

目的:以SemRep语义网及MeSH语义网的形式表达单篇论文的研究内容。方法:选取IF值≥3的10种医学期刊的20篇科学论文作为研究对象,利用自然语言处理工具SemRep及可视化分析工具Gephi构建SemRep语义网及MeSH语义网,并利用SPSS 23.0软件对数据进行统计分析。结果:MeSH语义网在表达单篇论文内容的全面性、准确性及易用性方面评分均值均高于SemRep语义网,且两种网络在准确性方面的评分差异具有统计学意义(P<0.05)。结论:相较于SemRep语义网,MeSH语义网更可作为一种表达单篇论文知识的方法,所提供的语义信息具备一定的文献挖掘潜力,能够为诸如文本挖掘、知识发现等研究提供新的方法和思路。

Objective To express the knowledge in a single using SemRep semantic network and Me SH semantic network. Methods Twenty scientific papers published in 10 medical journals served as the study object. The SemRep semantic network and Me SH semantic network were built using SemRep and Gephi. The data were statistically analyzed by SPSS 23.0.Results The average comprehensiveness,accuracy and useability scores of MeSH semantic network-expressed knowledge in a single paper were higher than in those of SemRep semantic network-expressed knowledge in a single paper. A significant difference was detected in the accuracy of MeSH semantic network-expressed and SemRep semantic network-expressed knowledge in a single paper was P<0.005.Conclusion Me SH semantic network is better than SemRep semantic network in expressing the knowledge in a single paper. The semantic information provided by MeSH semantic network and SemRep semantic network plays a certain potential role in mining literature and can thus provide new ideas and methods for the study on literature mining and knowledge discovery.