中文专利术语层次关系解析研究

发布时间：2018-03-11 17:01

本文选题：中文专利术语　切入点：位置加权　出处：《情报学报》2017年04期 　论文类型：期刊论文

【摘要】：对非结构化专利文本中的领域术语进行抽取以及语义关系的解析是挖掘蕴藏在专利文献中的丰富知识,并进行深入应用的前提。本文在领域专利术语有效抽取的基础上,探讨并实现较大规模术语层次关系的解析,构建了含有层次关系的领域知识本体。着重研究了基于位置加权的术语语义空间构建方法,基于主成分分析降维技术进行术语分布可视化以辅助聚类类目的确定方法以及术语层次关系结构中非重复性类目标签的抽取方法。本文工作尽可能实现了较大规模中文专利术语层次关系解析的自动化进行,为术语非层次关系解析以及基于术语语义关系解析的深入应用打下基础。
[Abstract]:The extraction of domain terms from unstructured patent texts and the analysis of semantic relations are the prerequisites for mining the rich knowledge contained in patent documents and for further application. This paper is based on the efficient extraction of domain patent terms. This paper discusses and realizes the analysis of the hierarchical relationship of terms in large scale, constructs the domain knowledge ontology with hierarchical relationship, and focuses on the construction method of semantic space of terms based on position weighting. Based on the principal component analysis (PCA) dimensionality reduction technique, the visualization of term distribution is carried out to assist in the determination of cluster items and the extraction method of non-repeatable target tags in the hierarchical relationship structure of terms. In this paper, as far as possible, the larger rules are realized. Automation of hierarchical relationship Analysis of Chinese Patent terms, It lays the foundation for the deep application of non-hierarchical relation parsing and semantic relationship parsing.
【作者单位】：南京大学信息管理学院;南京大学江苏省数据工程与知识服务重点实验室;
【基金】：江苏省自然科学基金项目“面向专利预警的中文本体学习研究”(BK20130587) 江苏省“333”工程项目“面向知识服务的中文本体学习研究”(BRA2015401) 国家社科重大招标项目“面向突发事件应急决策的快速响应情报体系研究”(13&ZD174)
【分类号】：G254;TP391.1

【相似文献】