情报科学 ›› 2024, Vol. 42 ›› Issue (12): 31-41.

• 理论研究 • 上一篇    下一篇

面向图书馆数据资源建设的智慧数据研究:内涵界定、技术应用 与研究启示

  

  • 出版日期:2024-12-05 发布日期:2025-06-27

  • Online:2024-12-05 Published:2025-06-27

摘要: 【目的/意义】全面梳理与综述智慧数据研究及应用情况,思考分析对文献图书馆数据资源建设可能带来的 影响,为学科研究发展提供参考借鉴。【方法/过程】通过系统梳理与计量分析国内外智慧数据重要相关文献,刻画 智慧数据发展态势与主题布局,对智慧数据的概念内涵、要素特征、组织理论模型、技术体系架构以及典型应用场 景进行深入阐述与分析。【结果/结论】智慧数据作为一种具有场景化、数据化、语义化与知识化的智能引擎,借鉴其 特征提出了图书馆在数据资源建设方面,建议“面向国家重大复杂科学问题的认知与决策场景、面向AI场域的多场 景、面向垂直领域场景以及面向产业创新全链条场景”建设场景化、专业化、关联化与知识化的数据知识资源。【创 新/局限】系统性梳理智慧数据的国内外研究现状,结合图书馆数据资源建设提出了相应的现实启示。

Abstract: 【Purpose/significance】Comprehensively sort out and review the research and application of smart data, think about and ana⁃ lyze the possible impact on the transformation and upgrading of document information work, and provide reference for the development of subject research.【Method/process】Through systematic combing and quantitative analysis of important relevant literature on smart data at home and abroad, the development trend and theme layout of smart data are described, and the conceptual connotation, ele⁃ ment characteristics, organizational theoretical models, technical system architecture, and typical application scenarios of smart data are in-depth expounded and analyzed【. Result/conclusion】Smart data, as an intelligent engine that is scenario-based, data-based, se⁃ mantic and knowledge-based, draws on its characteristics to propose the construction of data resources in the library, and recommends "facing the cognition and decision-making scenarios of major national complex scientific issues, and oriented "multiple scenarios in the AI field, scenarios for vertical fields, and scenarios for the whole chain of industrial innovation" to build scenario-based, profes⁃ sional, relevant and knowledge-based data knowledge resources.【Innovation/limitation】A review and analysis is conducted from the perspective of combining smart data and library data resource construction models, but the research literature may be missing, and the author's suggestions may be incomplete or even wrong.