Research On The Construction And Application Of Science Policy Knowledge Graph Based On Deep Learning

At present,the country attaches great importance to technological innovation.The government has issued a large number of technological policies to promote technological development and improve the supervision of technological enterprises.The companies need to obtain policy trends in time and make strategic adjustments in advance.However,science and technology policy items are complicated,and the companies need effective tools to obtain information related to their own interests from numerous policies.Google put forward the concept of knowledge graph in 2012,which has attracted widespread attention from academia and enterprises.The knowledge graph can integrate loosely structured data on the Internet and reorganize the structure of network data so that people can obtain information in a more effective way.Therefore,the knowledge graph can organize knowledge in a more efficient way and help companies quickly obtain effective information from a large number of policy texts.So,this paper combines the knowledge graph with science and technology policy,uses the knowledge graph to construct the science and technology policy knowledge network.Then realize the integration of science and technology policy information,and explores the application of knowledge graphs in practice with the help of the node search function of the graph database.The paper systematically sorts out the existing processes,key technologies and methods of the knowledge map construction.Then analyzes the relevant knowledge of the policy and the structure of the policy in combination with the status quo of policy research.This paper improves the classic seven-step method and builds an ontology database of science and technology policies.In the policy ontology,the type and industry category of the policy text are important information of the policy text.But the knowledge cannot be obtained directly from the text.Therefore,this paper chooses to use a deep neural network-based text multi-label classification technology instead of named entity recognition.Information is extracted.In addition,displaying the key content of the policy text in the knowledge graph can help companies quickly obtain the key content of the policy,so this paper chooses the text summary algorithm to extract the knowledge of the science and technology policy text.Combined with the analysis of existing research,this paper uses the Bert+BiLSTM-CRF model for text multi-label classification,and uses the TextRank algorithm to extract the key content of the science and technology policy text as the knowledge node of the knowledge graph.Then the Neo4j database is used to store the extracted knowledge units and draw the knowledge graph.This paper verifies the feasibility of constructing the process of science and technology policy knowledge map through practice.Based on the constructed knowledge map,the article explores the application of the science and technology policy knowledge map in policy services.This paper constructs the science and technology policy ontology,which provides a reference for the construction and application of the industry-wide policy knowledge graph.And the paper verifies the construction process of the science and technology policy knowledge map,explores the actual application scenarios of the knowledge map,and provides a reference for enterprises.It has certain significance for improving the efficiency of information communication between the government and enterprises and helping enterprises to quickly obtain key policy information.