The Construction And Application Research Of Academic Knowledge Graph Oriented To The Field Of LIS

With the development of the Internet,artificial intelligence and semantic web technology,knowledge graph has emerged.The knowledge graph is used to describe the entities that exist in the real world and the relationships between entities,and the entities and their relationships form a huge semantic web.Because of its powerful knowledge organization and knowledge representation capabilities,knowledge graphs are widely used in many fields,such as intelligent search,intelligent question answering systems,personalized recommendations.As a domain knowledge graph,the academic knowledge graph is oriented to academic data,such as journal papers,conference papers,dissertations,and scholar information on online academic resource websites.The academic knowledge graph can effectively organize these academic data in a visual way,and realize the transformation from data to information and then to knowledge.The author of the thesis presents the knowledge of the field learned by writing scientific literature,and the rest of the research users understand a research field by reading these scientific literature.Therefore,it is necessary to deeply dig and analyze the content of scientific literature and improve the efficiency of scientific literature utilization.The title,keywords and other fields involved in the scientific literature can well represent the scientific literature,so it is necessary to organize and mine these fields reasonably.As an effective knowledge representation method,the academic knowledge graph can realize the rational organization of the fields involved in scientific documents,and then realize the content mining of scientific documents,provide richer semantic information for the retrieval of scientific documents,and improve the use of scientific documents.Rate,to help scientific research users intuitively and accurately retrieve the required scientific literature.This article uses the construction technology and method of knowledge graph to construct an academic knowledge graph for the field of LIS.Use the papers of the CSSCI database in the field of LIS as the data source.The structured data of the selected papers include the title,author,year,and journal.The title can be spliced with the abstract into text data because of the rich content of the paper,so treat title and abstract as unstructured data.Structured data is used as a surface entity;for unstructured data,use the deep learning model Word2vec and BiLSTM+CRF to extract entities,identify research objects,research topics,and theoretical and technical semantic entities,and then perform entity alignment to build an academic knowledge map.And use the graph database Neo4j for storage and visual display.After constructing the academic knowledge map,three aspects of applied research on the academic knowledge graph will be carried out.The first application is document clustering.The academic knowledge graph realizes document association based on research object entities,document association based on research subject entities,and document association based on theoretical and technical entities.The three are relatively independent functions.Based on this function,the research object entity,the research subject entity and the theoretical and technical entity are integrated to realize the document association,that is,the K-Means clustering algorithm is used to cluster the documents based on the three types of entities,and the document association is realized.Analysis of cluster cluster topics formed after clustering;the second application is journal clustering,which is obtained from the association between journals in the academic knowledge map and their published papers and papers and their research object classes,research topic classes,and theoretical and technical entities In order to study the association between subject categories and theoretical and technical entities,use the graph database Neo4j to store and visualize this association,and use the K-Means clustering algorithm to analyze the difference of journals;the third application is the co-occurrence analysis of theoretical entities.There is an association relationship between a paper and one or more theoretical and technical entities in the academic knowledge graph.This association relationship is converted into an association relationship between a theoretical and technical entity and a theoretical and technical entity,that is,two theoretical and technical entities are jointly The co-occurrence relationship used in this paper is also stored and visualized using the Neo4j graph database to provide method-level guidance for scientific researchers’ paper writing.