Research And Implementation Of The Public Opinion Monitoring System For Judicial Domain

With the great efforts of our country to promote the disclosure of case information in relation to judicial authorities,the masses pay more attention to related events of Public Security Organs,Procuratorates and Courts.Although public opinions in network can promote judicial reform and progress,every wall has two sides,especially the pressure put on the judicial organs by vicious public opinion will interfere with the independence of judicial power.However,there is no contradiction between the judicial independence and the respect for public opinion.The support of public opinions monitoring technology can greatly alleviate the public relations pressure of judicial authorities and promote the judicial authorities to exercise their powers more justly and fairly.Public opinion monitoring is a complex system that crosses the social science and data science,and combines multiple technologies.Based on the actual demand,this thesis combines the technical research and system implementation of public opinion monitoring.This thesis introduces the implementation of a public opinion monitoring system for judicial domain,and does several related researches based on the massive real data collected by the system.This thesis has explored the application of a variety of machine learning algorithms and graph theory knowledge on structured data of public opinion.It includes quantitative and qualitative analysis of the propagation patterns of different types of events;prediction of the final propagation level based on the initial propagation state;prediction of the next-day propagation trend based on the current propagation state;and mining opinion leaders in the propagation network,etc.This thesis has also explored the application of multiple deep learning and natural language processing techniques to unstructured data of text type.It includes identification of judicial events from hot topics on the Internet;detecting user emotions in massive microblog comments;and mining event related elements in news reports,etc.The system has achieved good performance in many tasks,and has been used in the production environment.Relevant researches have high application value.