一种链生灾害复合风险智能评估预警方法及系统
By constructing a disaster assessment and early warning method with multiple triggering rules and interpretability mechanisms, the problem of single triggering rules and black box models in complex chain-like natural disasters is solved, achieving highly accurate and interpretable early warning results, and supporting the intelligent and refined development of disaster management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA UNIV OF GEOSCIENCES (WUHAN)
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-17
AI Technical Summary
Existing disaster assessment and early warning methods suffer from problems such as single triggering rules and black box models in complex chain natural disasters, resulting in a lack of scientific interpretability and credibility of early warning results, which affects the accuracy and credibility of disaster assessment.
A systematic framework of 'disaster chain scenario deconstruction—data-driven modeling—interpretability analysis—spatiotemporal coupling early warning' is adopted. Combining multiple triggering rules and interpretability mechanisms, a disaster susceptibility assessment model is constructed through the Stacking strategy. The SHAP method is used to reveal the internal discrimination logic of the model, construct disaster triggering patterns under multiple scenarios, and finally generate comprehensive early warning results through risk matrix coupling.
It significantly improves the accuracy, timeliness, and adaptability of early warning for complex chain disasters, provides an intelligent early warning paradigm with traceable mechanisms, adjustable rules, and reliable results, ensures the scientific nature and interpretability of early warning results, and supports the refined and intelligent development of disaster management.
Smart Images
Figure CN122414818A_ABST