The Influence Of Fans Effect And Sharing Behavior On Short Video Diffusion

With the development of mobile internet and communication infrastructure in recent years,many short video apps are spring out and short video is taking the place of texture contents to be the main content people comsume online.However,the research on the diffusion process of short video is very limited,most studies on short video focus on the features of short video or the application of the algotithm in short video apps to improve user experience.This thesis is taking Douyin,the largest domestic short video platform as an example to investigate the diffusion process of short videos using diffusion of innovation theroy.Contributions of this thesis are listed as follows.This thesis initially analyses the features of short video platform,and explores how the effect of fans,the social network structure as long as operating activities affect the diffusion process of short videos.This thesis then proposes a diffusion model that improves the classical diffusion of innovation model based on scale-free network and theory of fans and psychology.The empirical results shows that the proposed model outperforms the classical diffusion model and time series model in the aspect of fitting accuracy,.In addition,the scale-free network structure is fitting better the short video social platform research background than the fully-connected network assumption,and the degree of innovation affects the diffusion process much more than the effect of word-of-mouth caused by users'share action.At last,a forecasting model is proposed to predict the diffusion process of short videos.The obtained results are also better than the traditional time series prediction model.Based on the proposed model,some managerial implications are provided to short video platform operators.In summary,the proposed model fits the diffusion process of short video,and this work provides important theoretical and pracital contributions to the diffusion process on social platform like Douyin.The proposed model can be applied to real system and guide platforms content distribution and advertising,and also provide reference to other diffusion process studies on social platforms.