Research On The Construction And Application Of Literature Knowledge Graph Of Service-Oriented Manufacturing

As an important means of knowledge cognition,knowledge graph can analyze massive texts into structured entity-entity relationships and then represent knowledge system.At present,the general literature knowledge graph,represented by Microsoft Academic Graph and AMiner Academic Platform,provides technical support for literature retrieval and academic recommendation.However,the mining of proprietary knowledge for specific fields is insufficient and semantic information is scarce.Serviceoriented manufacturing is a new industry integrating manufacturing and service,which has the characteristics of fast development and strong crossdomain integration.How to use knowledge graph technology to build literature knowledge graph of service-oriented manufacturing and provide knowledge application in multiple scenarios is of great theoretical and practical significance for promoting the transformation and upgrading of China’s manufacturing industry and expanding the research of knowledge graph of industrial manufacturing.Based on the related research in the field of literature knowledge graph,a literature ontology model containing semantic information is designed,a service-oriented manufacturing academic literature knowledge graph based on Neo4j graph database is built,and three kinds of application scenarios are proposed.The main contributions of the research are as follows:(1)A literature knowledge ontology model including six entity types,twenty-six entity attributes and six relationships among entities is constructed.Compared with traditional literature ontology,literature subject entities with implicit semantics and domain direction entities representing research hotspots are added.(2)At the level of knowledge graph instance construction,the structural theme model is applied to refine the theme entities containing semantic information;Using TF-IDF algorithm to extract hot keywords as domain direction entities;Construct the literature influence index,and calculate the influence attribute value of each academic literature with the help of AHP and entropy method;Based on author co-occurrence analysis,the cooperative relationship between author entities is extracted.In addition,the Neo4j graph database is selected to store entities and entity relationships,so as to realize the visual display of knowledge graph of service-oriented manufacturing literature.(3)The knowledge graph of service-oriented manufacturing literature is applied to typical scenes such as domain knowledge entity and relationship inquiry,domain knowledge intelligent question and answer,domain research community discovery,etc.,so as to promote the government of Industry-University-Research to deeply understand the frontier knowledge of the domain with multiple subjects and provide knowledge support for the innovation and development of service-oriented manufacturing from all walks of life.On the basis of traditional knowledge entity extraction based on explicit knowledge,the knowledge graph of service-oriented manufacturing literature constructed in this paper expands the method of knowledge entity extraction with implicit semantics,and provides ontology construction ideas,multi-class knowledge entity extraction methods and field application examples for knowledge graph construction in emerging industries.In the future,besides theoretical knowledge sources,we can further expand the sources of enterprise practical knowledge,and compare the similarities and differences between academic cognition of service-oriented manufacturing and enterprise practical cognition.