Research On Key Online Information Mining And Warning Of Unconventional Emergency

In recent years, research on online information early warning of unconventionalemergencies has become a hot point in the areas of emergency management as well as“scenario-response” turns out to be the basic paradigm. Under this background, thispaper tries to theoretically establish a theory model of online information early warningof unconventional emergencies; researching spreading law of the online informationbased on the consideration of derivative information with the complex network theoryto structure key information mining model; based on both models, researching on earlywarning model and strategies to a reference for responding unconventionalemergencies.Deep research on theoretical key online information early warning model means:first, defining and discriminating the scenario both in decision making and inunconventional emergencies and their semantics as well as researching on the intensionand manifestation of unconventional emergencies online information to obtain onlineinformation originating from viewpoints posted by subjects such as netizens and mediaon forums or other network platform post-emergency and in its development, whichincludes solitary original information and massive derivative information; secondly,researching on the treatment of online information and key online information on thebasis of search the relationship between scenario and online information; at last, basedon both researching, making theoretical model architecture of unconventionalemergencies online information early warning strategies in three dimensions: the overallarchitecture of models, the one of spreading law of scenario based unconventionalemergencies online information, and the one of information based. The researchindicates that scenario is the combination of previous, present and future situations,which is partly similar to original and derivative information of the online one. Theoverall theoretical model architecture would be reasonable from scenario and onlineinformation, the spreading and evolution law of online information and the recognitionand mining of key information, to key information early warning.Deep research on the spreading and evolution of unconventional emergency online information means: first, during its process, the original trajectory varies with theupdating and superposition of derivative information. As a result, this paper highlightsconsidering the impact of derivative information to the original one, constructingderivative information based model of the spreading and evolution of unconventionalemergency online information, namely in considering the derivative information subjectto uniform, normal, chi-square and other continuous probability distribution toestablish the computational model and to conduct sampling calculation and simulation;secondly, it is the key point in describing and determining the spreading and evolutionlaw of unconventional emergency online information to effectively recognize keyinformation. Therefore, the fundamental trajectory can be seen to a certain extent byusing complex network model to construct unconventional emergency key informationmining model with mutual impact factor between derivative and online information.The results suggest that the computational model of the spreading and evolution ofunconventional emergency online information model with considering derivativeinformation can effectively recognize the key nodes (key online information); newmodel of key information mining with considering mutual factors can effectivelyrecognize the key nodes (key online information) in the law of it; to provide basis forearly warning.Deep research on early warning of key information based unconventionalemergency online information means: first, starting form key information, consideringnonlinear of the trajectory, imperfection, irregularity of information, and index hardlyquantified, and other characteristics, and combing through literature and practicalresearch, this paper design scientific early warning index system and withcomprehensive application of multi-level fuzzy evaluation And Trend analysis basedon the DEMATEL method to construct generalized qualitative and quantitativeemergency early warning model; secondly, taking into account governments, netizens,media and other Agents, this paper establishes Agent behavior models, causalitydiagram and flowchart, and through simulation, comes up with unconventionalemergency online information early warning strategies and guidance, which is of acertain significance.Aforementioned researches are in close connection with the reality ofunconventional emergency management needs, especially with the disposal of cases, with the integration of theories and the actual, providing help for the construction of "scenario-response " unconventional emergency management system.