Construction And Implement Of Company Knowledge Graph Integrating Leagal Documents

In recent years,with China’s continuous liberalization of the judicial field,a large number of judicial information has been released,such as the China Judicial Documents Network.As one of the most valuable judicial data,legal documents contain many enterprise information and become the focus of financial institutions.However,legal documents are characterized by many professional terms,strong logic and complex narrative content,which leads to the inefficiency of non-professional personnel who need to spend a lot of time and energy to extract document information.In addition,in the face of the problem that scattered and fragmented information in legal documents cannot constitute effective knowledge,this paper designs and implements the information extraction process of legal documents,and also constructs a company and legal documents knowledge graph(hereinafter referred to as CLKG)integrating legal document data on this basis to provide visual system services.The main content of this paper includes the following:Firstly,this paper designs and implements an scheme for extracting various types of information from legal documents.Based on the content and structure features of legal documents,the summarized rules and deep learning models are used to extract different types of information to ensure the overall extraction effect.The second is that the paper integrates data extracted from legal documents with existing enterprise business and relationship data,thereby constructing an enterprise knowledge graph containing new nodes and relationships.Unlike the enterprise knowledge graph that only includes enterprise and personal data,the knowledge graph in this paper can achieve more dimensional association mining,and providing a new direction for scenarios such as enterprise asset evaluation and risk prediction.Finally,this paper implements the research content in practice.In order to meet the business needs of both legal document information extraction and enterprise information management,an enterprise knowledge graph system based on CLKG has been designed and implemented to provide services such as legal document information extraction,enterprise information query,and graph relationship mining for business personnel.This paper is based on the professional process of software engineering.It ensures the feasibility of this project through preliminary project background and technical research,and analyzes the requirements of the project based on actual business scenarios.The four main functional modules of the system are determined:legal document information extraction,enterprise information query,enterprise graph query and enterprise case query.The detailed design content of the system is described through various professional charts such as E-R diagrams and time series diagrams to ensure the correct implementation of the system.Finally,in order to ensure that the system can be launched,this paper also conducted many system tests based on the above requirements.This paper has finally completed all the work to meet the system launch requirements,and the first version of the system has been successfully launched in the enterprise.This system can provide services such as legal document information extraction,enterprise information query,graph query,and case query.According to the constantly changing needs in the future,the system can also conduct in-depth mining based on the CLKG,to provide more graph query services.Based on the above work,the system has achieved two innovative results:one is a legal document information extraction scheme that combines rule extraction and deep learning models,realizing the ability to extract various types of enterprise information from legal documents.The second is the enterprise knowledge graph CLKG that added the "case"node and "enterprise-case" relationship.It integrates data extracted from legal document,and expands the graph,so that it provides possibilities for more practical applications.