Research On Key Technologies Of Scenario Information Acquisition And Analysis For Emergency

With the rapid development of science and technology and the constant progress of society, the life, work and social patterns of people have changed dramatically. At the same time, the natural and social environment that people face is becoming more and more complex. The interactions between scenario factors always cause that the local disturbance of some scenario factors may trigger a sudden unexpected event with serious social harmfulness. In the emergency management, the scenario information is the basis for the decision makers to analyze the development trend of the emergency, and the comprehensive scenario information is the fundamental guarantee for the scientific decision-making. So, how to fully and accurately obtain scenario information and how to automatically and quickly analyze scenario information, are key problems for any emergency management information system.For the past few years, semantic web and its related technologies are becoming more and more maturity, and have been widely used in the construction of emergency ontology model, the acquisition of scenario information, the analysis of scenario evolution, etc. However, there are many problems in the technologies of scenario information acquisition and analysis. First, the existing emergency ontology can not completely describe the complex relationships between events, scenarios, scenarios and events. Second, although linked open data has been widely used in the emergency management information system, there are many differences between the data model of data sets in linked open data, which limit the ability of people to get more comprehensive information. Third, during the emergency management, decision makers always have some priori knowledge, which is not considered in the existing conflict resolution methods for multi-source data. So, decision makers can not get more accurate scenario information. Fourth, the scenario evolutionary based on semantic web rules can only make a qualitative analysis of the emergency development, and can not help decision makers to analyze the scenario evolution quantitatively.Considering the above problems, this paper focuses on the key technologies of the scenario information acquisition and analysis in emergency. The main works and achievements are as follows:(1) By analyzing the characteristics of the emergencies and their scenario, we propose a emergency scenario information model, and define the syntax and semantics of the model with decription logic. The model is composed of event knowledge base, multiple scenario knowledge bases, resource knowledge base and association knowledge base. The model not only can describe the relationship between events, scenarios, scenarios and events, but also can limit or share scenario knowledge according to the relationship between events. Compared with the existing ontology model of emergency, the model can provide a powerful support for the acquisition and analysis of scenario information in emergency.(2) Considering the imperfection of the information query on linked open data, an approximate query method based on ontology alignment is proposed. Obviously, the results of ontology alignment have a great impact on the performance of information acquisition, so that we study the problems of the class and property matching between the data sets in linked open data. The class matching method we proposed utilizes the number of same objects to calculate the similarity between classes, and then get the matching class pairs. The property matching method we proposed makes use of the similarity between property functions to find the matching property pairs. The experimental results show that two matching methods both can achieve good results, which can help people get more comprehensive information from linked open data.(3) Basing on the characteristics of scenario information in emergency, a integration architecture for scenario information is proposed. In order to better handle the data conflicts, we present a framework to resovle the conflicts in heterogenous data by using prior knowledge, and adapt some conflict resolving methods to make them execute on the framework. The experimental results show that the heterogeneous data conflict resolution based on prior knowledge can get more accurate scenario information for the scientific decision-making, although prior knowledge in some cases may cause the measurement deviation of data source reliability.(4) According to the different roles of scenario in emergency evolution, we propose a scenario evolution model. In order to facilitate the implementation of scenario evolution model, we propose a rule langugage for scenario evolution by extending semantic web rule language from two aspects of syntax and semantics. The scenario evolution rule built on the rule language not only can qualitatively analyze the scenario evolution, but also can quantitatively obtain the probability of scenario happening.In this paper, the technologies introduced above are correlated with each other, and run through the whole process of information acquisition and analysis in the emergency management. The emergency scenario information model can provide guidance for the scenario information acquisition and analysis. By using the integration architecture for scenario information, the scenario information coming from linked open data will be integrated with the scenario information of other data sources. Then, the accurate and consistent information in the output of integration architecture will be deduced by the scenario evolution model. So, decision makers can continue to query linked open data according to the new scenario deduced from the scenario evolution model. Above all, the key technologies introduced in this paper can provide a powerful support for emergency management.