Research And Implementation Of Fake News Detection System Based On Machine Learning

With the progress of Internet technology,online social media websites play an important role in all aspects of today's life.Nearly three-quarters of people participate in at least one online social network,such as microblog and Zhihu.These types of online media provide their users with a lot of information to some extent.With the increasing use of social media,it is necessary to combat the spread of false information and reduce the dependence on such sources to obtain information.Social platforms have been under pressure to find effective ways to solve this problem,because users' interaction with false and unreliable news has accelerated its dissemination at the personal level.The spread of such misinformation has adversely affected people's perception of important activities,so the problem needs to be addressed quickly.The fake news detection system developed in this paper can detect fake news,curb the spread of fake news and reduce the impact of fake news on the people,so as to maintain social stability and build a social network platform with a good environment.The main research work of the text is as follows:(1)Through the survey,it is found that in social networks,users can directly express their views on news,so users' attitude towards news can be regarded as an important feature.At the same time,the author's emotional tendency is also contained in the news title.The author's attitude will have a direct impact on the viewer,which can not be ignored.Therefore,not only the features of user comments,but also the emotional features of news headlines are selected.Through the emotional analysis of the above characteristics,it can effectively provide support for news detection.In order to better detect false news,this paper also extracts text similarity features and user account credibility features.The features are input into the classifier for false news detection.Compared with a single text feature,it improves the accuracy of false news detection.(2)This system mainly designs and implements a false news detection system based on machine learning.The system mainly uses TF-IDF algorithm and cosine similarity to extract the similarity features of news text;Using the long-term and short-term memory network model integrated into the emotional dictionary,this paper analyzes the emotional characteristics of user comments and news headlines,extracts the emotional category characteristics of user comments and news headlines,and also extracts the credibility characteristics of user accounts.In the algorithm model,the improved naive Bayesian classifier is used to weight the extracted features,so as to realize the two classification of true and false news.(3)Combined with the current demand background object,this system analyzes the demand,determines the main functions that the system should have,and designs it in detail.The system mainly includes data crawling,news query,false news detection,visualization of analysis results,detection records and other functional modules.Finally,the article also shows the relevant implementation parts.