A government affair work order structured extraction method, device and equipment
By employing a dual-path retrieval and adaptive suggestion mechanism combining government public service knowledge graphs and vectorized indexes, the problem of dynamic adaptation to varying request types and responsibility boundaries in the structured extraction of government work orders is solved. This achieves high-accuracy and low-cost structured extraction, making it suitable for intelligent government public service systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INSPUR SOFTWARE TECH CO LTD
- Filing Date
- 2026-06-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for extracting structured government work orders suffer from problems such as fixed examples failing to cover diverse request types, poor dynamic adaptability of responsibility boundaries, and the inability of the example library to self-evolve, resulting in low extraction accuracy and high costs.
By constructing a knowledge graph and vectorized index for government public services, and employing a dual-path retrieval approach combining graph relationship matching and vector semantic similarity, along with adaptive prompts and a closed-loop update mechanism, the system dynamically recalls few sample examples and optimizes the example library, achieving high accuracy and low cost in structured extraction.
It significantly improves the accuracy and generalization of work order structure extraction, especially for low-frequency request categories, reduces knowledge base maintenance costs, and achieves an intelligent balance between reasoning efficiency and accuracy.
Smart Images
Figure CN122412499A_ABST