The Construction And Application Of Earthquake Disaster Scene Knowledge Graph Concerned On The Demands Of Multy-type Users

China is a country with frequent earthquake disasters.From 1998 to 2017,There were 120 earthquakes of magnitude 5.0 or above in China,including 7 earthquakes of magnitude 7 or above,resulting in a total of 486.16 million people affected,including 73,100 deaths and direct economy losses was 1.13 trillion yuan.The virtual scene of earthquake disaster can assist emergency decision-making by visually displaying earthquake information and disaster information.There are two problems on modeling earthquake disaster scene,firstly,little attention is paid to the cognitive ability and visual preference of multi-type users for earthquake disaster scene,which leads to the cognition difficulties of disaster information and low utilization of disaster scene,and secondly,because of the low degree of knowledge fusion in multiple domains,it is difficult to effectively link different information and make full use of different data,which results in low efficiency of scene modeling.This paper studies the use of knowledge graph in the construction of earthquake disaster scene and uses the feature that knowledge graph is a kind of semantic web,to effectively integrate the knowledge of virtual geographic environment modeling,earthquake disasters and multi-type users.By combining recommendation system method to take into account the semantic calculation concerned on multi-type users' visualization needs,we can construct the earthquake disaster scene conceptual model as needed.In addition,it assists in emergency response and is conducive to the rapid and effective implementation of earthquake and disaster reduction.The main research work and achievements in this paper are as follows:(1)Proposing a method to build earthquake disaster knowledge graph.Firstly,the earthquake disaster information,multi-user characteristics and virtual earthquake disaster scene characteristics are analyzed,and then the conception level and relationship composition of the corresponding ontology are designed.Finally,with the key links such as the characteristics analysis,knowledge extraction,knowledge fusion,and storage expression of earthquake disaster scene are systematically discussed,a method for constructing knowledge graph of earthquake disaster scenarios is formed.(2)Proposing an on-demand construction method for earthquake disaster scene.Based on the earthquake disaster knowledge graph,using personalized PageRank data mining algorithm,according to the directed graph network formed by the demand relationship between user nodes and disaster scene object nodes,the Personalized PageRank calculation is carried out,the semantic relevance value list which is closely related to user requirements is arranged,according to which,scene object nodes' value are from high to low to form a Top-N recommendation set.Finally,on-demand constructing the earthquake disaster scene conceptual model is realized.(3)The prototype system construction and application test.The earthquake disaster scene knowledge graph can be selectively displayed according to users' needs.Through the query and retrieval,you can view the general earthquake disaster scene knowledge graph and the subgraphs that reflect the knowledge of different fields;by running Personalized PageRank calculation,the user nodes with known visualization preferences,the importance of their demands are sorted,and the user nodes that do not give the visualization preferences,their probable visualization demand on earthquake scene are predicted.The actual experience and domain expert correction can verify the feasibility of the on-demand construction method about the conceptual model of earthquake disaster scene based on the Personalized PageRank proposed in this paper.