科技情报中的多源数据融合方法研究综述  

A Review of Research on Multi-Source Data Fusion Methods in Science and Technology Intelligence

作　　者：李芊芷 朱相丽[1,2] 李伟伟 Li Qianzhi;Zhu Xiangli;Li Weiwei(National Science Library,Chinese Academy of Sciences,Beijing 100190;Department of Information Resources Management,School of Economics and Management,University of the Chinese Academy of Sciences,Beijing 100190;National Innovation Institute of Defense Technology,Academy of Military Sciences,Beijing 100071)

机构地区：[1]中国科学院文献情报中心,北京100190 [2]中国科学院大学经济与管理学院信息资源管理系,北京100190 [3]军事科学院国防科技创新研究院,北京100071

出　　处：《情报杂志》2024年第9期157-165,共9页Journal of Intelligence

基　　金：国防科技战略先导计划(22-ZLXD-21-03-01-001-02的前沿“四性”技术识别研判及颠覆性技术清单研究拟制(第一阶段);22-ZLXD-21-03-01-001-04前沿科学技术等领域多源异构数据集构建)研究成果。

摘　　要：[研究目的]多源数据融合可通过交叉引证降低信息不一致和语义模糊等不确定因素,面对大数据和人工智能的普及导致的海量多源异构数据急剧增加,如何对其进行有效融合成为了科技情报领域研究的重难点之一。[研究方法]首先根据现有定义对多源数据的概念特征进行阐释;其次按照对文本中数据融合的深入程度,对多源数据融合方法进行系统梳理;最后分析了多源数据融合在不同科技情报场景中的应用和发挥的作用。[研究结论]从概念来看,多源数据融合具有内源性和外源性两种指向。从融合方法来看,现有常用方法主要包括三种:数据源的融合、结构关系融合以及语义融合,融合方法在实现从物理融合向化学融合转变的同时也在科技发展趋势研判、学术评价、需求探测等场景下得到广泛应用。未来研究需要通过实践操作细化异构加权的可行方法,以应用为导向加强对于融合方法的总结和归纳,探索形成可供复用的模型框架。[Research purpose]Multi source data fusion can reduce uncertainty factors such as information inconsistency and semantic ambiguity through cross referencing.Faced with the rapid increase in massive heterogeneous data from multiple sources caused by the popularity of big data and artificial intelligence,effective fusion has become one of the key challenges in the field of scientific and technological intelligence research.[Research method]Firstly,explain the conceptual characteristics of multi-source data based on existing definitions;Secondly,systematically sort out the multi-source data fusion methods based on the depth of data fusion in the text;Finally,this article analyzes the application and role of multi-source data fusion in different scientific and technological intelligence scenarios.[Research conclusion]From a conceptual perspective,multi-source data fusion has two directions:endogenous and exogenous.From the perspective of fusion methods,the existing commonly used methods mainly include three types:data source fusion,structural relationship fusion,and semantic fusion.Fusion methods not only achieve the transformation from physical fusion to chemical fusion,but also have been widely applied in the fields of technology development trend analysis,academic evaluation,and demand detection.Future research needs to refine feasible methods for heterogeneous weighting through practical operations,strengthen the summary and induction of fusion methods with an applicationoriented approach,and explore the formation of reusable model frameworks.

关 键 词：科技情报 多源数据 多模态数据 数据融合 语义融合 融合方法 

分 类 号：G350[文化科学—情报学]