Author(s): Jing Huang; Xingyan Wu; Huimin Wang

Linked Author(s):

Keywords: Bayesian network; Knowledge graph; Probability prediction; Rainstorm disaster chain mining

Abstract: In order to understand the evolution process of the rainstorm disaster chain and predict the secondary disasters and their impacts, it is necessary to develop disaster chain mining and prediction methods. Taking Pearl River Delta region as a study area, a knowledge graph for rainstorm disaster chains is constructed firstly, and rainstorm disaster chains are mined based on link coupling of knowledge graph. Rainstorm secondary disasters and their impacts in '8.29 Shenzhen Rainstorm' event are predicted based on Bayesian networks. The research results indicate that: a) The constructed knowledge graph for rainstorm disaster chains comprehensively depicts the rainstorm disasters and their impacts. b) The rainstorm disaster chain mining method based on link coupling reveals the evolutionary mechanism of the rainstorm disaster chain by coupling multiple independent events. c) The proposed Bayesian network prediction model for rainstorm disaster chains can effectively forecast secondary disasters and the impacts of disaster situations in the rainstorm disaster chain. The framework of rainstorm disaster chain mining and forecasting based on knowledge map can improve the accuracy and reliability of the model, enhancing the emergency response capability of rainstorm disaster chains.

DOI: https://doi.org/10.3850/iahr-hic2483430201-490

Year: 2024