Design And Implementation Of Social Media Data Collection And Fact Checking System

With the rapid development of the Internet,while it facilitates our lives,it also causes a large amount of false information to spread rapidly on social media.Facing the huge amount of information in social media,how to automatically identify the false content in it is a hot research direction in recent years.This paper mainly focuses on research in the field of fact verification.Different from general rumor detection work,fact verification work places more emphasis on specific evidence and interpretability,that is,it is necessary to judge the truth of a statement through evidence information such as evidence articles,knowledge bases,or knowledge graphs.fake.The core goal of this article is to design and implement an automated fact-checking system.The system will actively discover tweets with verification value in social media,and then collect web articles related to the statement in the tweet as auxiliary evidence,and then verify and analyze the statement based on the evidence to determine whether the statement is false information.The system is used to assist the staff of the rumor dispelling website to conduct fact verification.The research on algorithm analysis in this paper mainly has two points:one is to analyze whether a certain tweet has verification value.In this research,this paper first constructed a Chinese tweet verification value dataset,and then based on this dataset,built a model based on the BERT model combined with a downstream text classification model.The second is to conduct verification analysis and prediction based on network evidence for tweets with verification value,and judge the authenticity of the statement in the tweet.In this study,considering that there is a lot of redundant information in evidence articles in the real environment,it is necessary to pay special attention to the evidence selection link.Therefore,the model is constructed as a joint model architecture of the joint evidence selection task and the statement verification classification task,thereby promoting the real environment.Effective extraction of evidence information.In terms of system design and implementation,the application of this system needs to organize data collection,data analysis,data display,data storage and other links to interact with each other.In the specific study,this paper starts from the analysis of system requirements,and analyzes the functional and non-functional requirements respectively.Afterwards,based on the requirements analysis results,the architecture design work is carried out,and the system architecture design,function design,module outline design and data storage design are clarified.Finally,according to the design scheme,the specific realization of each module is carried out.After the system finally passed the software test,it was deployed and launched in the real environment.At present,the functional services of the system are normal and the operation is in good condition.