Knowledge Discovery Of Earthquake Emergency Based On Text Big Data

Earthquake is one of the major natural disasters that threaten the safety of human life and property.How to make effective decision quickly according to earthquake emergency knowledge and minimize the loss of life and property is the key of earthquake emergency.The method of building knowledge discovery model for earthquake emergency is one of the core scientific problems in the field of earthquake emergency knowledge.How to reduce the dependence on prior knowledge and support the knowledge discovery of earthquake emergency with a wide variety of data is crucial.At present,the knowledge field of earthquake emergency mainly focuses on knowledge acquisition,knowledge processing,and the data involved are mostly text data.It only focuses on the analysis of partial professional points in the field,and lacks the analysis of the whole field of earthquake emergency,and features a poor visual representation.Among the existing knowledge discovery methods,the statistical learning methods mainly rely on the limited knowledge of historical cases,while the methods based on machine learning and neural calculation require the prior knowledge of experts.The visual analysis methods tend to assist the knowledge representation.While this group of methods appear to be monotonous when directly used for knowledge discovery.Therefore,this paper proposes a knowledge discovery model for earthquake emergency based on big text data.First collect academic literatures related to earthquake emergency data and social media data sets.Then use CiteSpace analysis tools and formal concept analysis methods to extract high frequency keywords and inherent correlations,with word frequency correlations as the strength of the their relationship.Build a complex network of earthquake emergency knowledge to identify the communities in the network.Perform the search index analysis to obtain the public interest knowledge of spatial and temporal distribution characteristics of earthquake emergency response.Finally realize the earthquake emergency knowledge discovery and verify.This paper mainly includes the following two aspects:(1)A knowledge discovery model for earthquake emergency response based on big text data is proposed.Firstly,the paper introduces the source of big text data,which are the data sets of Chinese and English academic literatures and social media.The keywords of different data sets were extracted by the keyword extraction technology of text processing,and the correlation relations of corresponding keywords were established by cosine distance of included angle and formal concept analysis.Then,with the support of complex network theory,a keyword complex network was constructed,and the Louvain algorithm,which is most suitable for the network,was applied for the community identification.The modularity Q was used to extract reliable communities and the corresponding knowledge discovery analysis was carried out.Finally,some key words were introduced into Baidu index and Google index to analyze the public interest characteristics of the corresponding earthquake emergency knowledge,thus completing the construction and application of the entire earthquake emergency knowledge discovery model.(2)The findings of earthquake emergency knowledge were analyzed according to the model.Respectively on the academic literature,social media,search index knowledge discovery results corresponding analysis and cross comparison,clarifying the knowledge organization of earthquake emergency response under different data sources,found the implied correlation knowledge,find out the combination of multi-disciplinary professional knowledge,comprehensive knowledge for the earthquake emergency decision makers to provide decision support.