Method Of Emergency Key Attribute Recognition Based On Feature Network

After entering the 21st century, there have been some emergencies. Such as, Chinese SARS event, Wenchuan earthquake in Sichuan province, Japan’s Fukushima nuclear leak accident, Shanghai Bund trample disaster. The incidents which has a great harm to society, has aroused great concern in international community and domestic public, government, scholars. From the perspective of system theory, emergency have an intricate link with natural system and social system. Furthermore, the complexity and uncertainty of emergency lies in the sudden of emergency and information explosion. Information in emergency also increases the difficulty of emergency management. In this situation, the method of emergency decision making based on strict logical reasoning can’t meet the condition of time and s pace constraints. On the contrary, the method of emergency decision making based on a few key information solves the time and space restrictions. Thus, how to recognize the key attribute of emergency from massive information become one key problem in emergency response.According to current study, it usually ignores the reasoning rules in emergency evolution when it describes emergency situation and key attribute recognition. Combined with the theory of system science, knowledge science, management science and mathematics, this paper proposes the method of emergency key attribute recognition based on feature network. Mainly studies are as follow. First of all, based on knowledge element representation model, it proposes the object-attribute-relation model of emergency problem domain which analysis the logic relationship between event attributes and reasoning rules from three levels. From the perspective of knowledge element instantiation, it gives the use method of knowledge in object-attribute-relation model. Secondly, combined property knowledge and rule knowledge, it proposes the method of feature network structure which represents the development trend of the problem domain. The feature network is compose of event attribute as node and briefest structure of attributes in reasoning as directed edge. Then, from the perspective of emergency reasonable, it analysis the process of key attributes recognition. Combined the analysis of network centrality, it proposes the key attributes recognition method which based on the feature network. Finally, it proves the feasibility of method through a problem domain in Zhouqu debris flow.This paper which research from the view of knowledge organization is meaningful for emergency evolution and emergency decision. The object-attribute-relation model which based on element model implements a multi-dimensional organization of knowledge of things, reduces the difficulty of emergency knowledge in professional field for emergency decision-makers. The representation of emergency development trend based on feature network is more accordant to the characteristic of emergency complexity. It can enhance the accuracy of emergency development trend. The method of emergency key attribute recognition based on feature network adopt emergency management strategy which based on few key information. It can improve response efficiency of emergency for public sector.