档案资源建设2026年第40卷第5期《档案学研究》
场景牵引的高质量档案数据集建设路径——基于104个高质量数据集典型案例的扎根分析
Scenario-driven Path for Building High-quality Archival Datasets: A Grounded Analysis Based on 104 Typical Cases of High-quality Datasets
郑州航空工业管理学院 郑州 450046
摘要
面向高质量数据集建设与“人工智能+档案”持续推进背景,亟须探索高质量档案数据集建设路径。以国家数据局发布的104个高质量数据集典型案例为样本,采用扎根理论,提炼高质量档案数据集建设的一般逻辑,并结合档案数据的来源基础、载体形态、场景属性与治理要求进行领域适配。研究发现,高质量档案数据集建设可概括为“场景输入—生产机制—价值输出与反馈”的基本路径,业务赋能目标与现实数据缺口构成场景输入,指向档案业务应用、模型真实性校准、数据要素化利用等需求;围绕差异化场景,应形成由采集汇聚、工程加工、知识组织、可信治理、协同保障构成的生产机制,并经由模型训练、要素化运营与场景验证实现能力生成、价值转化与持续迭代。
关键词:高质量档案数据集数据治理数据要素档案数据要素
Abstract
Given the advancement of high-quality datasets and "AI+Archives", this study explores a path for building high-quality archival datasets. Based on 104 typical cases released by the National Data Administration, it uses grounded theory to identify the general logic of high-quality archival dataset construction and adapts it to archival data in terms of source, carrier form, scenario attributes, and governance requirements. The findings show a path of "scenario input-production mechanism-value output and feedback". Operation empowerment and data gaps constitute scenario input, pointing to archival applications, factual calibration of models, and data-element utilization. Data aggregation, engineering processing, knowledge organization, trustworthy governance, and collaborative support form the production mechanism, while model training, element-oriented operation, and scenario validation enable capability generation, value transformation, and continuous iteration.
Key words: high-quality archival dataset; data governance; data element; archival data element
引用格式
参考文献


