Research On Identification Of Health Rumors In Social Media Based On Deep Learnin

With the rapid rise of social media platforms,the spread of information is becoming more and more convenient.Due to the characteristics of Weibo’s strong interactivity and diversified topics,Weibo has become one of the most widely used social media in China,and health-related rumors can also spread quickly with the help of Weibo.Health rumors can spread widely on Weibo in a short period of time,which is easy to cause adverse effects and cause greater harm.Therefore,how to accurately and quickly identify health rumors in Weibo and reduce the harm caused by health rumors to the masses and society has become an urgent problem to be solved.Taking health rumors in Weibo as the research object and ELM theory as the support,this paper selects the characteristics of rumor recognition from three dimensions: Weibo text,emotional characteristics and user information,and combines deep learning methods to construct the rumor recognition model in this paper.The obtained Weibo rumor information was identified,and the traditional machine learning method SVM and random forest were used to conduct experimental analysis on the same dataset.The results show that the rumor recognition effect using deep learning method can achieve 99% accuracy,98% accuracy,100% recall,F1-score of 0.99 and AUC value of 0.99 based on the selected features supported by ELM theory,and the deep learning method performs best compared with SVM and random forest methods.It shows that the deep learning method adopted in this paper has better performance in rumor recognition.The simultaneous use of SVM and random forest methods can also obtain good accuracy,89.7% and 97.8%,respectively.It shows that the characteristics selected with the support of ELM theory can have a good rumor recognition effect.