Emergency Domain-driven Data Mining Middleware

With rapid development of society and economics, more and more emergency events have happened. However, how we can get actionable knowledge from dataset of emergency becomes more and more important to us in order to decrease emergency event and improve command level. There are many data analytical and mining tools or software in sub-domain in emergency domain. Emergency domain-based data mining middleware in order to meet the needs of sub-domain of emergency domain is built with domain engineering method.Domain-driven data mining combines data mining with domain special and domain expert in order that actionable knowledge can be mined form domain data sets. Emergency domain data mining is also to get actionable knowledge from emergency data sets, which can be used in many emergent management systems/platform in that many emergency events can be predicted and protected. In this paper, we proposal an emergency domain-driven data mining framework, emergency data set ontology and implement a emergency domain-driven data mining middleware. There are many component in the emergency domain-driven data mining framework includes: domain expert, domain knowledge and experiments, problems understanding and definition, problems constrained, data understanding, data preprocess, modeling, data mining result evaluation, post processing and knowledge delivery. The aim of data set ontology makes the data inputs standard. At last the paper also proposals an emergency ANN model, which is also validated in the emergency domain-driven data mining framework and implements in the emergency domain-driven data mining middleware.