Work

Research on the Classification Dimension Framework of High-Quality Data Sets for Large Models in the Emergency Management Industry

Year: 2026
Type: conference-paper
Abstract: To address the challenges such as inconsistent data standards, prominent information silos, and difficulties in integrating and sharing multi-source heterogeneous data in the high-quality dataset integration and sharing of the emergency management industry, this study, based on the mixed-line and plane classification method, constructed a four-dimensional classification framework for high-quality emergency management datasets. This framework covers four dimensions: information elements, business domains, data sources, and database types, and is further divided into classification labels from level one to level four, forming a clear and logically rigorous classification system. In the research, the IRP (High-Quality Dataset Planning) top-down method and the CGAF (China Government Architecture Framework) business method were combined to deepen and verify the classification system; through the IRP bottom-up review mechanism, the classification results were iteratively optimized in combination with actual business system data, and finally a refined classification directory covering multiple entities such as natural disasters, production safety accidents, emergency forces, and management objects was formed. The research results can provide standardized support for the unified aggregation, intelligent retrieval, sharing exchange, and comprehensive application of high-quality emergency management datasets, promoting the emergency management system to develop towards a data-driven, business-collaborative, and intelligent decision-making direction.
Language: English
Subfield: Safety Research
Open Access status: hybrid
OpenAlex ID: W7204994635
FWCI:
Cites: 8
Cited by:
Related to: 9