A temporal knowledge graphs prediction method for community gas risk

Abstract

A community gas risk prediction method based on temporal knowledge graphs is proposed to solve the complex community gas risk early warning problem. First, a community gas safety risk assessment indicator system is constructed based on the risk sources and factors influencing gas accidents. The entity and relationship features are extracted from the index system to construct a temporal knowledge graph of the community gas system. Then, a gas system risk prediction method based on the temporal knowledge graphs is proposed, which uses the RGCN algorithm to aggregate the information of neighboring nodes of the knowledge graphs in space and RNN to get the knowledge graphs information in temporal order and encodes and decodes it to make risk prediction based on the two kinds of information. Finally, the method’s effectiveness is verified by simulation under laboratory conditions based on a community in Beijing discarding historical data.