Statistical and semantic analysis of rumors in Chinese social media

In this study, we collect a large number of rumors from Sina Weibo, and analyze them from different perspectives. We study the influence of rumors from a quantitative perspective, as well as their beginning and ending. From a semantic analysis perspective, we manually classify the rumors into five categories based on the semantic information. Then, we train a classifier to classify the rumors automatically. From a timing perspective, we manually classify the rumors into four categories based on the forward trends. Then, we find the propagation tendency of rumors. Finally, we propose a rumor-recognition system, which combines machine intelligence and swarm intelligence.