Application Of Multivariate Heterogeneous Information Fusion Method In Court Execution Case Processing

In recent years,our country has been constantly exploring and making progress in the construction of "intelligent court".The two technologies most frequently used in the construction of the "intelligent court" are big data and big data,which are mainly used in the sharing of all kinds of information materials,and the other is artificial intelligence,which is mainly used in the trial.It carries out auxiliary examination of legal documents,automatic extraction of evidence materials of trial documents,etc.,which improves the application requirements of trial quality assessment,auxiliary trial,automatic sentencing,and precise division of cases.To achieve the development needs of the "smart court",the data in each system will be of various types,and due to different file formats,there are "data isolated islands",which makes it difficult to fully mine and utilize the data.In view of these problems,the work of this paper mainly includesThe main contents are as follows:Firstly,aiming at the problems of diversified data patterns,complex structure and data duplication and incompleteness after the fusion of various platforms,the rules are designed for transformation.A clustering method based on internal code sequence values is proposed to deeply detect and clean the duplicated data,and a distributed storage platform is constructed for distributed storage.Secondly,aiming at the problem that the data feature is not distinct and the source data of the knowledge body is huge and complex,which leads to the difficulty of data mining,define the rules of data transformation and construct the ontology knowledge base.Establish the mapping relationship between the business system data model and the defined comprehensive management platform data model,and put forward the combination similarity matching algorithm;Finally,aiming at the problem of data conflict caused by the inconsistency of multivariate data,the rules of voting,mutual verification,credibility and data source dependence are constructed,the Markov logic network model is established,and the voting perceptron algorithm is proposed.experiments are designed to verify the effectiveness of the above methods proposed in this paperThrough the research on the application methods of big data fusion,mining and analysis in the judicial field,this paper achieves the goal of improving the court to deal with different types of data effectively through the continuous study of new algorithms and new models.