Research On The Construction And Query Method Of Knowledge Graph In Coalmine Based On Neo4j

With the development of mine informatization and technology,the new generation of information technology represented by Internet of Things,cloud computing,mobile communications and artificial intelligence has been widely used in the domain of coalmine.Although the innovation and popularization of science and technology promotes the steady progress of coalmine construction,it is still difficult to solve the problem of how to manage and accurately query the complex and scattered data in the coalmine domain.Therefore,the knowledge graph technology is innovatively introduced into the coalmine domain,and the various types of data information and the relationship are visually described through the form of graphs,and the accuracy and efficiency of information query are optimized.This paper mainly builds the knowledge graph based on the knowledge of coalmine discipline,improves the knowledge storage method by using Neo4 j,and proposes a precise and fast intelligent query method,and develops the coalmine domain knowledge graph system to provide intelligent and visual services.The specific research contents are as follows:(1)Coalmine domain knowledge graph construction.The structural characteristics of coalmine subject knowledge are analyzed,and the model structure of knowledge base is proposed for the types of entities and relationships in the domain.Based on the mainstream technology,the knowledge graph construction method that meets the characteristics of the coalmine domain is improved.Then the Lattice-LSTM based on BiLSTM-CRF is introduced for knowledge extraction.Finally,in order to make up for the deficiency of relational database storage knowledge,the graph database Neo4 j is used to store the coalmine knowledge graph in the form of attribute graphs.The experimental results show that the Lattice-LSTM model can improve the entity recognition and word segmentation precision by using explicit word information,and complete the extraction of entity and relationships in the coalmine domain.(2)Parallel naive Bayes algorithm for knowledge query method based on spark is proposed.By analyzing the traditional knowledge base query method,it is a mechanical matching in the database by using the keyword of the entity name to be checked.This method is easy to cause deviations in query results by improper search terms.Therefore,Naive Bayesian classification algorithm based on Spark is proposedto improve the knowledge query method.First,predefine the expressions of common problems and their key words in the coalmine domain as training samples.;then the problem classification model is trained on Spark;finally,the Cypher query is generated from the feature and classification tags to search for answers from Neo4 j.The experimental results show that Naive Bayesian classification algorithm based on Spark can significantly improve the training efficiency of the model,and ensure the accuracy of problem classification,and improve the diversity and effectiveness of the knowledge query.(3)Based on requirements analysis and architecture design,the knowledge graph system in the coalmine domain using the Spring Boot framework to complete the system construction.the functions of knowledge management,knowledge query and knowledge visualization are accomplished.The example shows that the knowledge graph system can not only modification and supplement the knowledge of coalmine,but also accurately query and visualize the knowledge and visualize,which verifies the reliability and practicability of the system.