Research And Implementation Of Weibo News Rumor Detection Based On Deep Learning

With the development of Internet technology,the amount of information in the Internet is growing at an exponential rate,and it has become a huge information base.So that a large amount of unreliable information can be quickly and widely spread among the crowd.The proliferation of rumors on social media may make it difficult for people to pick out the information they need from a large amount of data,which in turn affects people's normal life.Especially in the face of public emergencies,rumors can be extremely destructive.Rumor detection is a subtask of text classification in natural language processing.The main purpose of rumor detection is to identify news and determine whether they are false news(rumors).In nowadays society,Weibo has become an inseparable part of people's daily life.Therefore,this paper uses Weibo as a platform to implement rumor detection.The purpose of this research is to mine the characteristics of the microblogging news texts,and propose a new method of rumor detection.The main work is as follows:1)We introduce the rumor detecting method based on user's attitude.In social media,users can express their views on news freely and directly,and the user's attitude towards the Weibo news can be a very important feature.Therefore,the algorithm divides the comments into two granularities of different sizes:words and comment sentences,and proposes a rumor detection algorithm in combination with two granularities.The comparison with classical rumor detection algorithm is carried out to verify the effectiveness of the algorithm we proposed.2)We study the sentiment classification algorithm based on Weibo news text.Most of rumors with strong emotions.Based on the classical text classification algorithm and the sentiment dictionary,this paper proposes a sentiment classification algorithm,which makes the algorithm model pay more attetion to the sentiment words.The experiment results validate the efficiency of proposed method.3)An algorithm that combines the above two results.This paper mainly considers the influence of text emotion and user attitude on rumor detection.Therefore,the results of the above two models are combined through pre-training.4)Design and implement of prototype system.This paper uses Weibo as a platform to implement prototype system.