Construction And Application Of Natural Disaster Emergency Knowledge Graph

Natural disasters occur frequently and pose a huge threat to China.People’s life safety and social and economic development are greatly threatened by natural disasters.Disaster prevention and mitigation has always been an eternal topic for human survival and development.Currently,technologies such as various sensors and remote sensing platforms are developing rapidly,and disaster-related data resources are becoming more and more abundant.However,only a small part of them are used and transformed into knowledge.Knowledge graph provides a new method of acquiring,storing,organizing,managing and displaying knowledge information.How to use the knowledge graph to obtain key knowledge in the emergency field from data,make the data generate wisdom,and complete the three-level transformation of "data-information-knowledge",has become a challenge facing the Institute of Intelligent Comprehensive Disaster Reduction.At present,related data has increased sharply in the field of natural disaster emergency response,while emergency knowledge is obviously lacking.There is a problem of insufficient "data-information-knowledge" transformation capabilities.Traditional natural disaster emergency response mainly relies on domain expert knowledge,and disaster emergency management is carried out with manual experience,lacking knowledge discovery and application methods based on existing data.This paper proposes the construction and application method of natural disaster emergency knowledge graph studies the unified description framework of the concept layer of knowledge graph,establishes a comprehensive ontology database in the field of natural disaster emergency response;studies the construction process of the knowledge graph data layer,establishes specific element examples and rich association relationships;On this basis,the application and update mechanism of the knowledge graph is proposed to realize knowledge retrieval,knowledge reasoning,knowledge update,and emergency task-oriented data and method recommendation.The main research contents include:(1)Construction of natural disaster emergency knowledge graph concept layer.Analyze the characteristics of the four core elements in the field of natural disaster emergency response: natural disaster events,disaster emergency tasks,disaster data,and methods,and use ontology theory to uniformly describe each element from top to bottom,design and build a comprehensive ontology database in the field,define the conceptual levels,the attributes and the semantic relationships of the four elements,define an accurate and structured conceptual framework,form a good conceptual level knowledge system,and provide a theoretical framework foundation for subsequent data layer construction.(2)Construction of natural disaster emergency knowledge graph data layer.In view of the wide distribution and diverse types of data resources,the bottom-up data layer construction process is studied,the data acquisition and text preprocessing methods are studied for unstructured data ubiquitous in the network,and the extraction methods of entities and relationships based on the concept layer ontology library are studied.This paper further proposes a knowledge fusion method based on entity similarity and a comprehensive knowledge storage method to realize the transformation from unstructured data resources to structured knowledge triples,forming a rich data foundation for the application of knowledge graphs.(3)Application and update method of natural disaster emergency knowledge graph.In order to effectively use the knowledge graph to support knowledge-based emergency decision-making,this paper studies methods of knowledge query and visualization,knowledge mining,reasoning.Facing the needs of disaster emergency missions,research on data recommendation methods to realize the rapid recommendation of remote sensing data and the application of comprehensive recommendation of data and methods.For the timeliness of information,an update mechanism based on the knowledge graph architecture is proposed to realize the dynamic update of the knowledge graph.(4)Take the common flood disasters in China as examples to carry out experimental verification and analysis.Carry out conceptual hierarchy and correlation decomposition of flood disaster events,emergency tasks,disaster data,and methods,and build a comprehensive ontology in the field of flood disaster emergency based on Protégé software design;obtain data from the National Disaster Reduction Network and the Chinese Academic Literature Search Platform website.Carry out the process of named entity recognition,relationship extraction and knowledge fusion.Based on the Neo4 j graph database,store the rich entities and association relationships in the flood disaster emergency knowledge graph.Use the knowledge graph query language to retrieve,display,relevance analysis,knowledge reasoning,use ontology rule reasoning and graph network analysis algorithms to discover available data and method entities for emergency tasks,and perform entity relationships in the knowledge graph Update the operation to verify the feasibility of the research method proposed in this article.The construction and application method of natural disaster emergency knowledge graph firstly based on the ontology theory to uniformly describe the four elements of natural disaster events,disaster emergency tasks,disaster data,and methods,and extract structured knowledge elements from multi-source data on this basis.At the application level,it proposes knowledge query,knowledge mining reasoning,ontology reasoning and graph network analysis-based emergency task-oriented data method recommendation,and explains the dynamic update method of the knowledge graph,and finally a flood disaster emergency knowledge graph is built to verify the validity of the proposed method.