Spatio-Temporal Knowledge Graph for Meteorological Risk Analysis
Jiahui Chen, Shaobo Zhong, Xingtong Ge, Weichao Li, Hanjiang Zhu, Ling Peng · 2021 IEEE 21st International Conference on Software Quality, Reliability and Security Companion (QRS-C) · 2021
The climate change issues are getting critical in recent years. Many institutions put more focus on the meteorological risk analysis. Reducing the harm of meteorological risks is of great significance to the fields of disaster prevention and mitigation, risk management and crisis response. Most recent approaches of meteorological risk analysis are data-driven. It is hardly for the effective reasoning using these methods because of the lack of knowledge. Therefore, knowledge mining from large-scale heterogeneous meteorological risk data becomes a critical issue of the meteorological risk analysis. In this paper, we propose a knowledge graph-based meteorological risk analysis framework with the core knowledge model named KG4MR (Knowledge Graph for Meteorological Risk). We introduce the modeling pipeline, knowledge discovery and corresponding spatiotemporal analytical grammar for the framework. Our framework can formally present the conceptual hierarchical relations between risky weather events and human activity events as well as element attributes. It can also discover semantic relations between elements. We demonstrate an example regarding Beijing 2022 Olympic Winter Games project in order to depict the availability of the proposed framework. It shows that the discovered domain knowledge can strongly support the reasoning in meteorological risk analysis.