Research On Patent Technology Value Evaluation Method Based On Knowledge Graph

With the continuous development of a knowledge-based economy,knowledge innovation plays an increasingly important role in social construction and economic development,and intellectual property rights have gradually become the core competitiveness.In this context,as one of the important manifestations of intellectual property rights,patents are increasingly valued by the state and enterprises.Therefore,there is an increasing demand for the valuation of patents,however,most of the existing valuation methods lack the mining of intrinsic information and correlations of patents,which will inevitably affect the valuation results to a certain extent.The advantages of the knowledge graph in information mining and data analysis can effectively solve the above problems,so this study proposes a patent technology value evaluation method based on the knowledge graph.Taking the technical value of patents as the entry point,an evaluation index system is constructed.The constructed indexes are used as the benchmark to design the structure of the knowledge graph.Select the enterprises in Sci-Tech Innovation Board as samples,identify the entities of their patent information,clarify the relationship between entities,and construct the patent knowledge graph.Design and implement the patent evaluation system,and provide evaluation services for users with the help of the patent knowledge graph to improve evaluation efficiency and reduce costs for users.The main work of this study is divided into the following 3 parts.Firstly,regarding the literature related to patent value evaluation,the factors affecting the value of patent technology are analyzed and elaborated from multiple perspectives.On this basis,the operation manual of the patent value analysis index system produced by the State Intellectual Property Office and the research results of many scholars are integrated,and 12 indicators that meet the needs of this topic are selected to constitute a complete evaluation index system.Then,we obtain semi-structured data about enterprises and patent information and unstructured data of patent text from Wind,China Patent Database,State Intellectual Property Office,etc.,apply the BERT-Bi LSTM-CRF model to fuse attention mechanism to extract entities from patent text,and calculate the similarity between entities by Word2 Vec algorithm to achieve fusion of similar entities.The relationship between various types of entities is clarified,and the ternary data is formatted and stored into the Neo4 j graph database,which constitutes the patent knowledge graph.Finally,for the target patent provided by users,the constructed patent knowledge graph is used to retrieve the patent group belonging to the same technical field with a similar subject matter.Based on the information data of this patent group,the CRITIC weighting method is used to assign weights to each index,and then the technical value of the target patent is evaluated and an evaluation report is generated.A patent technology value evaluation system is designed and developed to facilitate user operation.This system adopts B/S architecture,the front-end is based on the Vue framework,combined with D3.js to realize the visual display of knowledge,and the back-end is based on the Spring Boot framework.The front-end and back-end realize JSON data exchange through the fetch interface.The database uses My SQL and Neo4 j storage modes comprehensively.