Research And Implementation Of Case-oriented Legal Knowledge Search Syste

The intelligence of the judicial field is constantly advancing,and legal provisions serve as guidelines for regulating the results of judicial trials.Nowadays,research on the intelligence of legal provisions is receiving increasing attention.When searching for knowledge of articles of law,currently users can only obtain the inherent static knowledge of articles of law,unable to quickly view the association relationship between articles of law and cases,and unable to understand the relevant facts and reasons for citing articles of law in specific cases.Therefore,research is conducted on how to quickly and accurately identify the data knowledge hidden in legal articles of case.It has important research significance and application value to clearly display the correlation information between law cases.In order to solve the above problems,based on the judgment documents data of judicial criminal cases and criminal law regulations,this paper adopts the method of machine reading understanding to study the method of case plot recognition related to articles of law,and displays the correlation relationship between criminal law articles and criminal cases in the form of knowledge graph.Finally,based on these research results,this paper implements a case-oriented legal knowledge search system.This system can not only help users quickly view the association relationship between the law and the case,but also let users understand the association case and the reason for citing the law in specific cases,namely,the knowledge of the law oriented to the case,so as to achieve a more intelligent and accurate search for the knowledge of the law.The main work of this paper is as follows:(1)This paper proposes a method based on machine reading comprehension model to identify the case plot related to the law.In this method,the law is regarded as a question,the text of the case fragment and the punishment fragment considered by the court is taken as the context,and the case plot related to the law is taken as the answer.The machine reading comprehension model based on MacBERT-BiLSTM was used to extract the answer fragments corresponding to the questions in the context,and the training and testing were carried out on 2000 question and answer pairs.The experimental results show that the F1 value and EM value of this model are90.05% and 88.09%,respectively,showing excellent performance.(2)Research on the Construction Method of Article-Case Temporal Knowledge Graph: This paper uses the data of criminal law articles and criminal cases,and constructs the temporal knowledge graph of article-case according to the three determining factors of target article,the circumstances of related cases and related cases.The knowledge map clearly shows the correlation between the law and the case,and introduces the time attribute of the relationship between nodes to explain the historical changes of the law,so as to design a new form of knowledge display of the law.(3)Design and implementation of case-oriented law knowledge search system:This paper develops a case-oriented law knowledge search system based on the plot identification method of law related cases and the construction of law-case sequential knowledge map.The system is developed by the Flask and Vue.js framework of the WEB application framework.Through demand analysis and software design,the system is built and displayed,and its practicability are verified.