论文部分内容阅读
[目的/意义]近年来关联数据数量飞速增长,伴随而来的数据垃圾、语义与语法错误等低质量现象屡见不鲜,数据集质量问题受到普遍关注。全面分析关联数据质量评价的研究内容与方法,为质量控制与评价体系的构建提供参考;[方法/过程]系统梳理国外关联数据质量相关的研究动态,总结评价方法的演变过程,归纳与分析各种评价类型的内容、模式、特点与优缺点,在此基础上进行文献综述;[结果/结论]关联数据的质量评价类型主要分为基于单一指标的评价、基于多维指标的评价与质量评价指标体系。质量控制与评价体系的研究将向指标程度完整、公认度高、可行性强的方向寻求发展。
[Purpose / Significance] In recent years, the number of related data has been rapidly increasing. Low-quality phenomena such as data trash, semantic and grammatical errors are commonplace. Data quality problems have received widespread attention. To provide a reference for the construction of quality control and evaluation system; [Method / process] systematically combing the research related to the quality of foreign related data, summarizing the evolution process of evaluation methods, summarizing and analyzing The content, pattern, characteristics and advantages and disadvantages of the evaluation types are summarized and literature review is conducted on the basis. [Result / Conclusion] The quality evaluation of the related data is mainly divided into the evaluation based on single index, the evaluation based on multi-dimensional index and the quality evaluation index system. The research on quality control and evaluation system will seek development in the direction of complete index, high recognition and feasibility.