CED: Credible Early Detection of Social Media Rumors

Rumors spread dramatically fast through online social media services, and people are exploring methods to detect rumors automatically. Existing typically learn semantic representations of all reposts a rumor candidate for prediction. However, it is crucial efficiently as early possible before they cause severe disruption, which has not been well addressed by previous works. In this paper, we present novel detection model, Credible Early Detection (CED). By regarding sequence, the proposed model will seek an point-in-time making credible We conduct experiments on three real-world datasets, results demonstrate that our can remarkably reduce time span prediction more than 85 percent, with better accuracy performance state-of-the-art baselines.