Research On The Construction Method Of Laws And Regulations Knowledge Graph

With the gradual improvement of people’s living standards,the legal awareness of the public is gradually enhanced,and the demand for legal knowledge is increasingly strong.The traditional legal consultation business relies on the lawyer’s answer,but in China,due to the small number of practicing lawyers,it is difficult to solve people’s legal problems in time.The purpose of this paper is to use natural language processing technology and legal expertise to propose a method that can automatically generate knowledge map including regulatory content,so that people can search the required regulatory knowledge faster.Because most of the legislative languages are based on the programming language,based on the characteristics of the regulatory language,this paper constructs the regulatory knowledge map.The main research contents are as follows:(1)We analyzed the logical structure of laws and regulations and the phenomenon of legal stylized language in the text of laws and regulations,and build the ontology,defined the role types and relationship types of legal terms in the laws and regulations knowledge graph.(2)In the process of knowledge extraction,it is divided into two parts:legal terms role recognition and relationship extraction.In the aspect of legal terms role recognition,we use a model BERT-BiLSTM-CRF which based on BERT pre-trainning language model.The experimental results show that this method has better performance than the traditional machine learning method and other deep learning network models in the control group,and achieves the best results in Fl-measure,precision rate and recall rate.(3)This paper designs a method that can automatically generate the knowledge map of laws and regulations,and displays the knowledge graph with neo4j graph database as the carrier.