The Research Of Emergency Case With Public Opinion Crawling,Classification And Semantic Retrieval

Contemporary news reports and related research articles about emergencies are explosive growth.Facing the massive emergency cases and public opinion,in order to extract events attributes and mine events correlation,this paper presents a model of emergency case and its public opinion which abstract climb,classification,retrieval and other work to templates.The model,design and implementation of the configurable requirements of data crawling,extraction and classification are carried out.This paper finished the following work.First,this paper puts forward the emergency case and the public opinion ontology model.The model is based on the ABC ontology model as upper ontology,and combine the data crawling,extraction,classification in vertical search engines and abstract the configurable template,implements the configurable model of case and public opinion crawl,case extraction and case classification.Second,this paper propose a design scheme of the process of emergency case with public opinion and semantic retrieval platform.For the feature and structure of emergency case and public opinion,we devise an overall platform framework as well as the platform resource transfer and data structure.According to the module partition,we make a series of design like data flow design,data structure design and function design for data crawl module,data extraction module,data classification module,case library module and case semantic retrieval module respectively.Third,this paper implement process of emergency case with public opinion and semantic retrieval platform.By Xinhua net news RSS feeds example,China's analysis emergency network news feature entity,instance of Lanzhou University case bank.We introduces the data crawl process,data extraction process,data classification process,case library and semantic retrieval functions of the platform from three aspects that is RSS feeds news,news features and case data base.This paper contribute the emergency case and public opinion model and implement configurable model of data crawling,data extraction and data classification.And we validate versatility of the model though the realization on different data sources,and provide an effective platform support for events attributes extracting and events correlation mining.