A Study About User Portrait Of Users Participating In Bullet Screen Interaction In Online Video Community

With the development of Internet information technology and the expansion of social networks,in the context of the increasing enrichment of pan-entertainment content,various communities around the user's personalized hobbies are blooming,with the result that the information interaction between has become increasingly frequent and close.We can find that people have pursued more Real-time,social and fun way during the process of user information interaction.Among these online interest communities,the emerging bullet-screen video community is a typical representative of the real-time interaction characteristics of users.Because of its unique interactive atmosphere,it has received an acclaim by many people.The Bullet-screen is used as a video real-time information interaction form for users.This type of interaction provides a broader space for users,meanwhile all kinds of junk Bullet-screen s have appeared which hampers the sustained development of community.Therefore,the research on the Bullet-screen users and their behavior characteristics has certain practical value.This article selects Bullet-screen as the entry point and focus on the users involved in the Bullet-screen interaction.It starts with the characteristics and concepts of the Bullet-screen,which reviews substantial relevant researches about information behavior of users in the network community and user portrait.From the researches about Information Interaction Behavior and user portrait in various field,especially in the field of library and information.Based on those researches,this article uses statistical analysis and human behavioral dynamics to describe the behavior of Bullet-screen information interaction by means of reptiles,trying to grasp the distribution of the characteristics of the users involved in the Bullet-screen interaction.This study found that the users involved in the Bullet-screen interaction didn't have significant differences in gender distribution,but level distribution was more concentrated,mainly ranging from three to five,which reflects users interacted in Bullet-screen are armed with high community stickiness and activity Besides we can understand the relationship between user characteristics and the characteristics of the barrage interaction behavior;Moreover,user level and followings are related to the number of Bullet-screen,With the help of human dynamics theory,it is found that the natural time intervals of two users who publish the Bullet-screen in the same video obey the power law distribution,and all accord with "strong matrix weak memory",which indicates that the user's Bullet-screen interaction behavior is consistent with "long-term silence,short-term outbreak".Finally,the machine learning algorithm is used to aggregate the users of different behavioral characteristics into three categories according to the number of Bullet-screen,the length of the Bullet-screen and the time interval.After portrait analysis,the three types of users can be described as "rational discussion user","Bullet-screen leading type" and "general Public type".Personalized service suggestions are proposed from five aspects:personalized search,accurate recommendation,resource integration,precise operation,and reference decision-making.It provides relevant suggestions and references for the development direction of the video community.