Research On Information Text Retrieval And Automatic Classification Of Earthquake Emergency Command System

Earthquakes in China are not only high frequency,lagest intensity,but also widely distributed.And the social economic loss caused by earthquakes ranked the first palace among natural disasters.Therefore,it is the key point to reduce the social economic loss caused by the earthquake,that the researchers of earthquake emergency concentrate on.Speak to earthquake emergency information management,there are 3main differences in China compared with the developed countries:(1)the information acquisition method is relatively single;(2)the information transmission is not sufficient and accurate with fragments and distortions;(3)the information acquisition speed is slow with diverse output and storage source.Therefore,in the consideration of the complicated mass emergency information uploaded from multi-channels and the rapid development of the information retrieval in the intelligent field,The thesis collected emergency data in the earthquakes of the Lushan Ms7.0 on April 20,2013,Kangding Ms6.3 on November 22,2014,Jiuzhaigou Ms7.0 on August 8,2017,and Guangyuan Qingchuan Ms5.4 on September 30,2017,and made the following analysis and research works:(1)Introduced the work systems and contents of the earthquake emergency in China in details,and analyzed the characteristics of the multi-channels of the emergency information.The information sources are divided into the field information source and the non-site according to the spatial attribution of the information output.And the information is sub-divided according to the channels and users combined the service objects of the emergency information.(2)Based on the results of earthquake emergency information classification from the experts and scholars in recent years,aiming at the high efficient and convenient information service,The information is divided into 3 categories including basic background informationbefore the earthquake(the natural disasters of the earthquake area and the background information of the social foundation),information during the earthquake(the property of the earthquake itself and the direct and indirect loss information caused by earthquake),and information of the rescue(emergency rescue action information).(3)By studying the text information retrieval,a high efficient and convenient classification method is structured in the consideration of difficulty in manual classification and the requirement of the classification.Compared the 3 basic retrieval models of information retrieval,this thesis carried out the automatic classification of the earthquake emergency information according to the traditional Boolean model(keyword classification)with the TF-IDF as the weight formula.(4)The automatic classification of multi-channels emergency information is carried out with the "keyword classification" method.The main works include: 1)information classification.The collected emergency data is classified according to the established classification standard;2)process of text structuralization.All information is collected in a text document(.txt file),and then the information is classified statistic analyzed according to the characteristic or property of a certain word;3)Keyword base is established according to selection of keywords.The keyword base is from the words of top 20 in frequency and with corresponding TF-IDF values.4)Taking the automatic classification is carried out in computer language,ranked in 5 intervals according to keyword frequency,through maximum matching of the Boolean model.After the information structuralization,the pre-earthquake basic background information includes the selected keywords of population,precipitation,GDP etc.The earthquake disaster information includes the disaster area,the earthquake condition,the aftershock etc.Thepost-earthquake emergency rescue information includes bands,search and rescue,medical equipment etc.From the selected results,the key words of all kinds of information are recognizable.The processed information promises logic and convenience of the information and its further processing.The automatic classification of the earthquake emergency information improves the utilization rate of information,and saves the time of the processing and providing information,which promises the quick and efficient services in the emergency and command.