Research Of User Flow On Gifting In Live Streaming And Prediction Model Construction

With the development of live streaming for many years,pan-entertainment live streaming has become an important part of the digital economy with strong social and emotional attributes.As a business model,gifting is the main revenue source of panentertainment live streaming platforms,and gifting is also known as virtual gift consumption behavior,which refers to an act that audiences make virtual gift purchases according to their own wishes after watching the performances of streamers.As one of the new forms of consumption,the live streaming economy is supported by national policies.However,with sound regulations,slowing user growth and intensifying competition pattern,panentertainment live streaming companies are facing the current situation of declining revenue.Meanwhile,in the academic field,the current research on live streaming is mostly focused on the product purchase intention of e-commerce live streaming users,and there are still relatively few studies focusing on the users’ gifting.In this context,this paper focused on the core research object of user gifting and explored the impact of user flow on gifting and predicted user repeated gifting from the perspective of Flow Theory and Customer Lifetime Value Theory in pan-entertainment live streaming.The first research question in this paper is the impact of user flow on their gifting.Panentertainment live streaming provides users with a comprehensive immersive spatiotemporal atmosphere,which in turn may change user behavior.Based on the classical Flow Theory in online consumption research,this paper investigated the influence of users’ flow in different subjects(platforms vs streamers)on their gifting from the users’ perspective,and explored the moderating effect of relation-based live streaming and task-based live streaming on this influence path.This paper selected Huya,a leading livestreaming platform in China,and collected user data from Huya through Python.Firstly,the BTM topic model was applied to extract topics from the danmu of ten users,and the texts were classified into four types of topics,namely emotional feedback,user interaction,cognitive interaction and self-disclosure,and it was found that the content of emotional feedback and user interaction accounted for the most and the content of self-disclosure accounted for the least in the users’ danmu.Subsequently,this paper used SPSS and Stata to conduct statistical analysis and hypothesis testing,and the results showed that:(1)Both users’ streamer flow and platform flow positively affected users’ gifting.(2)In terms of streamer flow,its impact on gifting was greater in relation-based live streaming compared to task-based live streaming.(3)As far as platform flow is concerned,it only had a significant positive effect on gift-giving of taskbased live streaming users,and had no significant effect on gifting of relation-based live streaming users.For the hypotheses that were not tested,possible explanations were given in the discussion section of this paper.The second research question in this paper is a study on the prediction of repetitive gifting.In this paper,the Customer Lifetime Value Theory in was introduced,assuming users’ repetitive gift-giving has individual inertia,and the prediction model was built by adding user flow as a covariate based on the classical BG/NBD model.In order to test the model effect,this paper collected consumption data from 1070 users of Huya platform to extract their behavioral characteristics for prediction,and used Mean Absolute Error(MAE),Mean Square Error(MSE)and correlation analysis to compare the prediction results.The results showed that the user’s repeated gifting model based on BG/NBD model can achieve the prediction well,and the model had the best prediction effect when the user’s streamer flow and platform flow were added to the model as covariates at the same time.In summary,this paper investigated user gifting and repetitive gifting layer by layer,explored the influence of user flow in different subjects on their gift-giving,and provided a prediction model of users’ repetitive gifting.There were three contributions of this paper:(1)This study understood the textual content shared in danmu through topic mining and identified the information needs of live streaming users.(2)A model of the influencing factors of user gifting was built,which expanded the research perspective of Flow Theory.(3)A more accurate prediction model of repetitive gifting was built.This paper enriched the research content and used more cutting-edge research methods in the field of livestreaming,and the findings of this paper provided good practical insights for pan-entertainment live streaming platforms,where companies should pay attention to the role of user flow experience on users’ gifting,and enhanced users flow by improving platform mechanisms and cultivating streamer skills.