Summary of Research on Network Rumor Detection

The research on network rumor detection is a classification problem of information which is to classify information into the credible and the distrust in accordance with the authenticity of itself. The main research rout is to construct the information set and select several features from text characteristics, user characteristics, propagation characteristics, emotion features etc, and then select a appropriate classifier to divides the information as rumors or non rumor. Rumor is from makers, the author believes that the psychological and behavior characteristics of makers are also important bases for rumor detection, and it's the research trend in the future. The existing rumor detection method all judges whether the information is a rumor, which is too simple. Therefore, this paper argues that, we should choose more appropriate detection method and integrate various characteristics to calculate the probability that the information is a rumor.