Mining Of Abnormal Relationship Among The Parties Involved Based On Information Extraction

In the existing research,the algorithm of social network community detection is usually used to mine criminal gangs,for example,based on social network analysis or cluster analysis to mine criminal gangs.These two methods are more suitable for the discovery of criminal gangs with a small amount of data.For larger data sets and unknown criminal gangs,it is impossible to effectively discover and mine.This article uses the method of combining the attributes of the persons involved,the construction of character portraits and the detection of abnormal subgraphs to find the abnormal relationship between the characters from the text,that is,the abnormal behavior group.The main work is as follows:First of all,in view of the incompleteness of the attribute extraction of the involved persons,this article conducts ontology modeling of the involved parties and uses the AE＿PI extraction framework to extract attribute information.From the four aspects of place-things-people-behavior,the person involved in the case is modeled on the ontology,and the portrait of the person involved is constructed.Then use the AE＿PI extraction framework proposed in this article,which is based on the combination of the BERT pre-training model and the rule extraction model,to extract the attributes of the person involved from the legal judgment document data.Secondly,in order to be able to mine potential criminal gangs from a large amount of text data,this article proposes a framework for mining abnormal relationships among persons involved based on information extraction.The framework combines the character attribute extraction algorithm and the abnormal subgraph detection algorithm,and through the realization of a series of operators and data processing,finally outputs an abnormal character relationship network,which is represented as a abnormal behavior group.Finally,in order to prove the validity and feasibility of the framework for mining abnormal relationships among persons involved,this article conducts an empirical analysis of the framework based on the data from the judgment documents of the crime of provoking disturbances.The development of an ontology-based extraction system is used to extract the information of the persons involved in the judgment document data,build an attribute network based on the attributes of hometown and crime location,and conduct case analysis.In summary,this article combines location,event,person,and behavior to model the person involved in the case.At the same time,it uses the extraction algorithm of this article to extract the attributes of the person from the text data,and combines the person field with the abnormal sub-graph detection field.The method is to find the abnormal relationship between the persons involved in the case from the data of the legal judgment documents,that is,the abnormal behavior group.From the existing data in the judgment documents,unearth unknown groups of people involved in abnormal behavior.