期刊文献+

基于多源数据融合的科技决策需求主题识别研究 被引量:21

Topic Identification of Scientific and Technical Decision-making Demands Based on Multi-source Data Fusion
下载PDF
导出
摘要 大数据环境下,单一数据对科技决策支持的服务不充分、不全面,传统科技决策需求的获取方式较为被动,面对情报用户的决策需求愈加复杂的状况,增加了对用户情报需求描述、情报需求解读与情报服务的难度。为协同利用多源数据,使不同来源的信息相互补充,主动探测情报用户需求,提出一种基于多源数据融合的科技决策需求主题识别方法,以我国科技部机构用户需求主题识别为例,综合多源文本数据结合主题强度分析确定科技部机构用户的重点关注领域主题,确立主题属性,对需求主题在各个主题属性文本中进行词向量计算,从而主动识别出细粒度更高的机构用户科技决策需求主题。通过以科技部机构用户需求主题识别为例,融合多源数据相互补充印证,实现了主动捕获与探测用户的情报需求,从而对用户需求的掌握更加客观、合理、有效,丰富了情报服务的模式,为情报服务变被动为主动提供新理念、新方法。 In the big data environment,single-source data are not sufficient and comprehensive in the service of scientific and technical(S&T) decision-making support,and the access ways of traditional S&T demand are rather passive.The situation in which decision-making demand of information user is becoming increasingly complicated has increased the difficulty of user’s information demand description,illustration,and intelligence service.In order to synergistically utilize multi-source data,so that information from different sources can complement each other,and to actively detect the demand of information user,this paper proposes a topic identification method of S&T decision-making demands based on multi-source data fusion.Taking the topic identification of Ministry of Science and Technology(MOST) institutional users’ demands as an example,the paper synthesizes multi-source text data and combines topic intensity to analyze and determine the topics of their concerned areas,establish the topic attributes,and perform word vector calculation on demand topics in the texts of each topic,thus proactively identifying the fine-grained topics.Through the topic detection of MOST users’ demands and the complementation and verification of multi-source data fusion,the paper achieves the active capture and detection of users’ intelligence demands,so as to grasp the demands of users more objectively,reasonably,and effectively,enrich the modes of intelligence services,and provide new ideas and methods to the transformation from passive intelligence services to positive intelligence services.
作者 周群 化柏林 Zhou Qun
出处 《情报理论与实践》 CSSCI 北大核心 2019年第3期107-113,共7页 Information Studies:Theory & Application
基金 国家社会科学基金面上项目"基于多源数据融合的情报用户需求探测研究"的成果之一 项目编号:17BTQ066
关键词 多源数据融合 科技决策 用户需求 主题识别 multi-source data fusion S&T decision-making user demand topic identification
  • 相关文献

参考文献15

二级参考文献273

共引文献607

同被引文献285

引证文献21

二级引证文献136

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部