Research On Automatic Construction And Application Methods Of Event Networks For Drug-related Case

In recent years,with the advancement of the construction of "smart courts",courts have accumulated a large amount of judicial data centered on judgment documents,which play a key role in the application of judicial intelligence assistance.The case facts in the judgment documents are the main basis for judges to make judicial judgments.How to effectively obtain the case characteristics in the case facts is of great value to the application of judicial intelligence assistance.The event network is constructed by mining the case characteristics composed of criminal event trigger word and related elements in the case,which can support smart judicial applications such as crime prediction and case information completion.Most of the existing case representation methods construct a case knowledge graph by extracting event features from the case.However,these methods can only obtain limited case features and ignore the connection between the case features.Therefore,this paper takes the judgment documents of drug-related cases as the breakthrough point,and studies how to construct an event network of drug-related cases to analyze drug-related cases,so as to solve the problem of low efficiency of manual analysis of drug-related cases.On this basis,the neural network model is used to learn the law of the case in the event network,and it is applied to the task of completing the case information to provide a reference for the analysis of the integrity of the case.This paper mainly does the following work:(1)Aiming at the analysis of the facts of drug-related cases,this paper proposes to construct an event network to represent the facts of drug-related cases.Taking criminal events as nodes and event relationships as edges,the automatic construction of the drugrelated case event network is realized through the process of case text extraction,criminal event extraction,event merging and event relationship identification.Crime fact information,in order to support the research of judicial intelligent auxiliary application based on case analysis.(2)On the basis of constructing the event network of drug-related cases,this paper proposes a case information completion method based on the event network.The cases newly received by the court may be unreviewed,and there may be a problem of lack of case information.Court staff need to analyze and judge the integrity of the newly received cases.This paper uses a case information completion method that combines event sequence features and event network features to predict missing case information.By learning the event representation of drug-related cases and predicting the missing criminal events in the event network,the case information is completed.The accuracy rate of 63.56% was achieved on the data set of adjudication documents of drug cases.