关键词: 微博, 谣言检测, 谣言事件, 卷积神经网络

Abstract: Manual detection rumors often require a lot of manpower and material resources, and there will be a long test delay. At present, the existing rumor detection model is generally based on the rumors of the content, user attributes, transmission of artificial structural features. Feature building is the key, but it can not avoid the existence of one-sided, waste of human and other phenomena. In order to solve this problem, this paper presents a rumor detection model based on convolution neural network, vectorizes the rumor events in microblog, mines the deep features of text through the learning and training of hidden layer of convolution neural network, avoids the problem of feature construction, and can find those features that are not easily found, and produce better results. The experimental results show that this method can accurately identify rumors, and it is better than the comparison algorithm in accuracy rate, accuracy rate and F1 score index.

Key words: microblog, rumors detection, rumor events, convolution neural network