Research On Probabilistic Rule Mining And Hybrid Reasoning In Emergency Scenarios

When an emergency occurs,it is important to implement a quick and effective re sponse in disaster relief.However,to implement a quick and effective response,predic ting the evolutionary trend based on the current emergency scenario is a crucial factor.Therefore,rapid and accurate prediction of the developing trend of emergencies has b ecome the basis for the effective implementation of response activities.Reasoning on emergencies needs to be supported by the basic methods.The reaso ning process often involves the expertise of multiple disciplines,and also faces proble ms of mining,management and application of massive information,rules,cases and k nowledge.Therefore,in order to improve the feasibility and reliability of the reasonin g on emergency evolution,it's necessary to take reference,integration and improveme nt of traditional reasoning methods to adapt to the complexity of emergencies.Therefo re,in the process of the reasoning of emergency scenario evolution,how to combine case-based reasoning(CBR)and rule-based reasoning(RBR)to make up for their resp ective deficiencies and to construct an effective hybrid reasoning model has become a n important scientific problem that deserves a closer look.Based on the basic knowledge-element model,this paper formally describes the e mergency scenarios,emergency cases and emergency reasoning rules,depicts the interr elationship between the emergency scenarios,cases and rules,and interprets the comm on features of emergency scenario evolution.On this basis,by combining CBR with R BR,it proposes a hybrid reasoning model of emergency scenario evolution based on t he knowledge-element model,and elaborates its executive processes and methods.The model can help decision-makers scientifically predict the developing trend of emergen cies and provide decision support for formulating emergency response plans,thereby r educing the losses caused by emergency incidents.In addition,in order to deal with the problem that emergency reasoning rules ma y be absent in the process of hybrid reasoning of emergency scenario evolution,this paper constructs a rough-set-based probabilistic rule mining model of emergency scenar ios.First,the scenario information of emergency cases is organized into a decision tab le based on the rough set theory.Secondly,genetic algorithm is applied to reduce the att ributes of the decision table of emergency cases' scenarios.Then,by implementing probabilistic rule mining algorithm,the probabilistic rules of emergency cases' scenarios are o btained.It is of great significance for understanding the potential correlation between elements in emergency cases,exploring the potential rules hidden behind historical em ergency cases,and supporting the reasoning of emergency scenario evolution.Finally,the practical application process of the hybrid reasoning model was empl oyed in Along hill's forest fire,which occurred in Greater Khingan Mountains.Also,the obtained reasoning values were compared with the actual values,verifying the effe ctiveness and scientificity of the model.In the process of the hybrid reasoning,the pr obabilistic rule sets could be obtained by the probabilistic rule mining model of emerg ency cases' scenarios.Thus,once forest fires occur in the area,they can provide deci sion-making support for the attribution of related resources,such as personnels and eq uipments,which is of great significance for formulating forest fire warning and respon se plans in the Greater Khingan Mountains.