- 期刊
A Survey of Evaluating Credibility based on Deep Learning in Online Social Networks
摘要
The rapid development of online social networks makes the dissemination of information more rapid and convenient, but the dissemination of false information may have a negative impact on the majority of Internet users, resulting in difficulties in platform management, social unrest, and even the development of the country in serious cases. Therefore, it is significant to evaluate the credibility of the content in online social networks. In the era full of data, models such as recurrent neural networks and convolutional neural networks in deep learning technology have excellent data mining ability. Through feature mining of existing data, we can predict the credibility of new data. To solve the challenge of credibility evaluation, this paper first defines the credibility evaluation of online social networks, then systematically summarizes the credibility evaluation methods in recent ten years, and expounds on three kinds of credibility evaluation methods. Secondly, the performance of the classified reliability evaluation technology is evaluated by using the relevant evaluation indexes. Finally, through the analysis and summary of the existing work, this paper further puts forward the possible research direction of online social network credibility evaluation technology in the future.