The Design And Implementation Of Data-Quality Driven Interpretability Analysis System For Judgement Documents

In recent years,our country has made significant progress in the construction of smart courts.However,there still exsit many problems that need to be solved urgently.Judicial data,as an indispensable part,plays an important role in the construction of smart courts.The judgment document is the carrier for the court to perform its judicial duties and demonstrate judicial justice.Due to the continually increasing trend in the civil and commercial cases as well as the growing transparent of judicial justice,the quantity and quality of judgment documents have attracted more and more attention.In a large number of judgment documents,it is an important challenge about how to control the quality of documents and make good use of the information conveyed by these documents remains to be studied.To alleviate the problem of excessive length and redundant information of judgment documents,this paper proposes the interpretability analysis technology of judgment documents based on data quality which aims to improve data value density.It combines advanced technology like text abstract and dependency parsing with judicial rules to measure and improve the quality of judgment documents by fact model and multi-dimensional measurement system about data/information quality.The multidimensional measurement system covers six dimensions of standardization,completeness,readability,responsiveness,delay,and richness,with a total of 24 indexes.Judicial rules and advanced technology ensure the accuracy and utilization of judgment documents respectively.In order to reduce the cost of users and maximize access to document information,the analysis service described in this paper is the core of the technology.Based on the low-cost and high-efficiency operations,this paper designs a fact model and builds a multi-dimensional measurement system about data/information quality espectively from the perspective of single file and batch file.After automatically processing and analyzing data via NLP,the tasks present the interpretability analysis reports for judgment documents.Meanwhile,problem localization of documents is detected and suggestions for repair are provided.This paper designs a system,including data interaction module,measurement analysis module,and quality improvement module.It provides a user-friendly interactive interface to help users easily analyze data and perceive representative information.