突破性创新早期识别与弱信号分析综述  

A Review of Early Recognition of Breakthrough Innovations and the Weak Signal Analysis

作　　者：刘亚辉 许海云 Liu Yahui;Xu Haiyun(Chengdu Library of Chinese Academy of Sciences,Chengdu 610041;Department of Library,Information and Archives Management,School of Economics and Management,University of Chinese Academy of Sciences,Bejing 100190;Business School,Shandong University of Technology,Zibo 255000;Institute of Scientifie and Technical Information of China,Beijing 100038)

机构地区：[1]中国科学院成都文献情报中心,成都610041 [2]中国科学院大学经济与管理学院图书情报与档案管理系,北京100190 [3]山东理工大学管理学院,淄博255000 [4]中国科学技术信息研究所,北京100038

出　　处：《图书情报工作》2021年第4期89-101,共13页Library and Information Service

基　　金：国家自然科学基金项目“基于科学-技术主题关联分析的创新演化路径识别方法研究”(项目编号:71704170);国家重点研发计划项目课题“颠覆性技术地平线扫描系统”(项目编号:2019YFA0707202-01);国家自然科学基金项目“科技关联视角下新兴技术弱信号扫描预判方法研究”(项目编号:72074014)研究成果之一。

摘　　要：[目的/意义]通过比较分析不同的突破性创新识别方法,总结现有方法存在的问题,将弱信号引入突破性创新研究的识别中,重点关注突破性创新早期的各类弱信号,尤其是弱关联关系分析,以期实现早期预判。[方法/过程]首先,通过调研现有的识别方法,提炼当前存在的主要问题,指出研究弱信号的必要性。之后,从不同的学科角度介绍弱信号的内涵及表征形式,对其特征进行概括,梳理弱信号的几种识别方法。最后介绍弱关系分析的内涵及应用,提出借鉴多元关系融合算法模型可以实现多种弱关系的有效融合,获取更明确的信息。[结果/结论]突破性创新的识别研究中受关注最多是文献间的引用关系、主题词之间的语义关系等强关系数据,而弱关系蕴含着更多元化的信息,加强弱关系分析可以实现对突破性创新研究的早期预判。未来研究需要寻求有效捕捉弱关联的方法,注重主题的动态演化规律,如利用高阶网络模型分析有效弱信号,提高突破性创新早期识别的准确性。[Purpose/significance]This paper has summarized disadvantages of various recognition methods of breakthrough researches by comparing and analyzing these methods.Then this study has introduced weak signals into the identification of radical innovation,focusing on various types of weak signal(especially weak ties)w ith a view to achieving early prediction.[Method/process]Firstly,by analyzing the existing recognition methods,this paper summarized the main problems and pointed out the necessity of weak signal research.T hen,this study introduced the concept and representation of weak signal from different disciplines summarized the characteristics of weak signal and combed several methods of weak signal identification.Finally,this paper introduced the connotation and application of weak tie s,and proposed that the algorithm model that integrates multiple relationships can realize the effective fusion of various weak ties and obtain more clear directional information.[Result/conclusion]In the frontier research,most attention is paid to strong ties such as reference relations between literatures and semantic relations between topic terms.However,weak ties contain more diverse information,and strengthening the analysis of them can enable early recognition of breakthrough research frontier.In the future,it is necessary to seek effective methods to capture weak ties and further dig out the evolution of the theme such as:using high-order network models to analyze weak ties to improve the accuracy of early recognition.

关 键 词：突破性创新 早期识别 弱信号 弱关系 关系融合 

分 类 号：C250.2[社会学]