Feature-aware emergency decision recommendation with multi-graph convolutional networks

Emergency decision-making for unexpected events at civil transportation airports involves dynamic situations, time constraints and information scarcity. Rapid and effective post-emergency decision implementation remains a critical concern for relevant departments and emergency academia. With the development of graph neural networks, their application in recommendation systems has become a hot research topic. To improve recommendation quality, more and more researchers model data as an information network with two types of nodes: users and items. This approach facilitates more accurate knowledge discovery. However, existing studies often do not fully utilize the comprehensive structural and rich semantic information within the network. Therefore, this paper proposes an emergency recommendation model incorporating multiple types of entity-present learning networks to alleviate the pressure on commanders in responding to accidents. The experimental dataset was obtained on a simulation platform, and the experimental results show that our method has made significant improvements compared to advanced recommendation methods. Further research has demonstrated the effectiveness of the emergency decision-making method approach this paper proposed.