语义增强的科技创新内容表征研究  

Research on the Representation of Technical Innovation Content with Enhanced Semantics

作　　者：白如江[1] 祝娜[1] 王效岳[1] 

机构地区：[1]山东理工大学科技信息研究所,山东淄博255049

出　　处：《情报理论与实践》2016年第3期73-79,共7页Information Studies:Theory & Application

基　　金：教育部人文社会科学研究基金青年项目"长句检索中信息查询扩展研究"(项目编号:12YJC870001);山东理工大学人文社会科学发展基金项目的研究成果之一

摘　　要：[目的/意义]随着计算机信息处理技术以及文本数据挖掘技术的发展,研究人员开始利用语义分析技术深入分析科技文献文本数据,识别出科技文献中的科技创新内容,以期为科技创新和科技决策提供支持和帮助。[方法/过程]文章通过分析科技创新内容结构分布特征,以句子为最小标引粒度,利用Keygraph算法抽取出科技文献摘要中的关键词进行科技创新内容特征选择,基于SVM的语义角色标注技术完成科技创新内容的语义表征。[结果/结论]实验结果表明,语义增强的科技创新表征方法可以基本实现科技创新内容的语义标引。[ Purpose/significance ] With the development of computer information processing technology and text mining tech- nology, researchers use semantic analysis method to analyze scientific literature data and identify technical innovation content in the scientific literature in order to provide support and assistance for technical innovation and technical decision-making. [ Method/ process ] By analyzing the distribution structure of technical innovation content, the paper takes sentence as the minimum indexing granularity, and uses Keygraph algorithm to extract the keywords in the abstract of scientific literature for the selection of character- istics of technical innovation content. Based on the semantic role labeling technique of SVM, the paper accomplishes the semantic representation of technical innovation content. [ Result/conclusion] The experimental results show that technical innovation repre- sentation method with enhanced semantic can basically achieve semantic indexing of technical innovation content.

关 键 词：语义增强 科技创新 算法 支持向量机 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] G254[自动化与计算机技术—计算机科学与技术]