学术论文大纲中关键术语抽取方法研究

发布时间：2018-01-17 13:46

本文关键词：学术论文大纲中关键术语抽取方法研究　出处：《现代图书情报技术》2014年03期 　论文类型：期刊论文

【摘要】：【目的】针对学术论文大纲内容精炼、层次性的特点,研究从中抽取重要且具有实质意义术语的方法。【方法】结合语言学规则和术语词典从大纲各级标题中识别出候选术语集,然后根据术语间的句法依存关系计算tf-idf,并利用大纲结构量化术语层级特征,最后结合tf-idf与层级特征对候选术语进行排名,选择出关键术语。【结果】实验证明,该方法的候选术语识别F值达到89.57%,术语选择F值达到36.89%。【局限】采用的术语抽取规则不完备,且tf-idf计算过程中的权值设置仅使用经验值,导致未能达到最优效果。【结论】该方法能有效抽取大纲中的关键术语,适用于层级结构中的关键术语抽取。
[Abstract]:[objective] to aim at the characteristics of refinement and hierarchy of the outline of academic papers. This paper studies the methods of extracting important and meaningful terms from them. [methods] combining linguistic rules and terminology dictionaries, candidate terminology sets are identified from syllabus titles at all levels. Then tf-idf is calculated according to the syntactic dependency between terms, and the hierarchical features of the terms are quantified by the outline structure. Finally, the candidate terms are ranked with tf-idf and hierarchical features. [results] experiments show that the candidate term recognition F value of this method is 89.57, and the term selection F value is 36.89. The term extraction rules used in this method are not complete. In the process of tf-idf calculation, only the empirical value is used, which leads to the failure to achieve the optimal results. [conclusion] this method can extract the key terms in the outline effectively. It is suitable for the extraction of key terms in hierarchy.
【作者单位】：中国科学院国家科学图书馆;中国科学院大学;
【基金】：国家科技支撑计划子课题“基于文献知识网络的领域学术关系研究与示范”(项目编号:2011BAH10B06-04)的研究成果之一
【分类号】：TP391.1
【正文快照】： 1引言关键术语(Keyphrase)是指反映文章主旨的词或短语,常在自动文摘、信息检索、文档聚类和自动问答等自然语言处理系统中表示文档[1-3]。关键术语抽取(Keyphrase Extraction)是从大量的文本中识别关键术语的过程,抽取结果对自然语言处理应用有重要影响。传统的抽取方法主要

【参考文献】