Modeling video viewing and sharing behaviors in online social networks

Both video viewing and sharing behaviors in online social networks are of great importance for optimizing traffic engineering and understanding the information diffusion mechanism. However, few models have been proposed to characterize these two behaviors together. In this paper, we first collect the sitewide viewing and sharing statistics of videos in a popular OSN in China, and propose a new model, VSM, to capture the temporal dynamics of video viewing and sharing behaviors during the diffusion process. Specifically, our model can handle the external influence and periodicity properly. The experiments based on the collected dataset demonstrate our VSM can outperform other alternative models in terms of explanatory power and prediction accuracy.