Design And Implementation Of Semantic Analysis And Visualization System For Judicial Dispute Data

Tremendous development has been made in informatization,but the existing traditional government data processing system has no way to meet the requirements of the times.The data is massive and complicated which cannot be effectively utilized.It caused a large amount of meaningful data accumulated and cannot make positive influences for government decision-making and people's livelihood issues.In order to solve this problem,we use several technical means to separate the social difficulties and hot issues buried in the massive data from the complicated and redundant information,and specifically analyze the disputes at a certain place and a certain time to find out the social problems deeply buried.Meanwhile,we use visual analysis technics to transform the data above to interactive visual charts,which helps municipal staff obtain information more effectively.In order to achieve the above objectives,this paper designs a method based on weighted TF-IDF algorithm to realize the classification of judicial dispute text,and implements a set of judicial dispute data visual analysis system based on this method.The system aims to realize automatic classification of judicial short text data such as dispute mediation records,and uses visual analysis technology to present,which helps users to understand the current social dispute situation intuitively and accurately while ensuring data accuracy.Discover hidden social hotspot issues.The system starts with the original dispute mediation data,through data annotation,model training,automatic classification,and visualization.This article divides the above steps into three parts.(1)Natural language processing.The corpus of real dispute text of a certain city is processed through six steps including: data cleaning,Chinese word segmentation,semiautomatic text annotation,TF-IDF word weight calculation,weighted word weight TFIDFc calculation,model training.And in the end a well-trained classification model will be obtained,which can be used to devide the corpus into several specific kinds of dispute.The results are used for the visual analysis work.(2)Visual analysis component.Based on visual analysis methods such as tree diagram,wordcloud,and sparkline,the data of judicial disputes are integrated and optimized,and different rules are used to display multi-level tree diagrams.Meanwhile,wordcloud and sparkline are integrated to make dispute mediation data.Rendering can be done in multiple dimensions of time and space.(3)Dispute mediation data visual analysis system.The above classified data,word cloud data and other information are used for the visual analysis component,and the front and back ends are integrated through the B/S framework to realize a complete visual analysis system.The system visualizes dispute resolution data in the form of visual charts to help users capture social hotspots intuitively and accurately.