Research And Implementation Of Knowledge Graph Technology For Emergency Supplies Management

In the era of web 2.0,with many applications sinking into more specialized fields,the characteristics of highly specialized information and discrete distribution are prominent,so it is more necessary to fully explore the correlation between information.Knowledge graph technology can help solve this problem.It uses the tool of graph database to represent complex unstructured relationships between entities,sorts professional domain knowledge into a scientific and orderly knowledge system,and makes finding and managing related knowledge more convenient,fast and efficient.precise.Emergency material management is a very important part of the public affairs emergency management system.It is the basic consideration and backup force to do a good job of prevention before an emergency occurs,to do a good job of scheduling when it occurs,and to do a good job of aftermath.Therefore,aiming at the management of emergency materials,this paper uses knowledge graph technology to construct a knowledge system of accident investigation reports and emergency plans for natural disaster accidents and production safety accidents.Based on the knowledge graph construction technology in the professional field,this paper focuses on the entity name recognition technology and relation extraction technology of domain knowledge.Investigated the processing methods of entity naming recognition tasks used by existing research in professional fields,and proposed a solution that combines domain knowledge rules,BERT pre-training model and BiLSTM-CRF model for named entity recognition,and the effect has been significantly improved.This paper proposes a two-layer joint extraction model of entity relationship based on multi-information fusion and dependency analysis to solve the task of entity relationship extraction.Aiming at the problem of unclear entity relationship status in relevant documents in the field of emergency materials management,two solutions are proposed.The preset entity relationship is transformed into a relationship classification task.Supervised dependency parsing and semantic role tagging methods for relation extraction.After the triplet set is extracted,a knowledge map in the field of emergency materials management is constructed,including three components:emergency materials,natural disaster accidents and hazardous chemical accidents.Finally,this paper constructs a knowledge map in the field of emergency materials management through the processes of ontology construction,knowledge acquisition,knowledge fusion and knowledge storage.Design and implement a knowledge map system for emergency materials management based on knowledge map,which can help users to query and manage emergency materials efficiently and intuitively,specifically in the knowledge map of emergency materials,knowledge map of natural disasters and accidents in safety production.Graph query entity nodes and relationships.It provides a reference for the construction and application of knowledge graphs in similar professional fields.