关键词: 科学数据管理, FAIR原则, 国家科学数据中心, 文本挖掘

Abstract: [Purpose/Significance] This paper explores the FAIRification characteristics of data policy of National Scientific Data Center in China, aiming to provide a preliminary reference of data management policy formulation and work optimization for them. [Design/Methodology] This paper comprehensively used the methods of network research and text mining. The content mining software, KH Coder, was employed to conduct quantitative text analysis of 79 data policies from 20 data centers. Through analyzing the frequency of FAIR principle appear in these policy texts and the words with high similarity used in FAIR principle of these policy texts, we revealed the attention difference and semantic feature of FAIR principle in different data centers and different types of policy texts. [Findings/Conclusion] The results show that the data policies of data centers have reflected some FAIR principles, but the attention to each principle is not balanced. Different types of data policies focus on different aspects of the FAIR principle, and the commonality lies in the findable principle and interoperable principle and a strong emphasis is given to metadata. [Originality/Value] This paper suggests that in the development of data policy, National Scientific Data Centers should highlight the role of "metadata" in data lifecycle management, promote the construction of data policy system driven by "data value-added" and appropriately introduce the FAIR principle based on the scientific data management practice in China.

Key words: Scientific data management, FAIR principles, National Scientific Data Center, Text mining