Research On The Structural Characteristics And Influential Factors Of The Interactive Network Of Bilibili Interest-based Community

The development of video socializing has deepened,and the form of interest-based communities has been constantly enriched.As a video platform with prominent features of interest-related,bilibili has a high proportion of users of Generation Z+ and has consumption potential.The motivation and results of interaction among members of interest-related community need to be further studied.This study takes the interest-related community on bilibili as the research object,and innovatively incorporates the topic and emotion of interactive comments into the research model based on the theory of interactive ritual chain.Specifically,from the perspective of social networks,the exponential random graph model is used to explore the generation of interactive networks,and the structured topic model combined with Doc2Vec algorithm and deep learning methods are used to measure the focus and emotion of community members.The specific content and research results are as follows:(1)Construct a model of the structural characteristics and influencing factors of the interactive network of the bilibili interest-related community.The model is based on the theory of interactive ritual chain,and uses interactive ritual reinforcement feedback to represent the interactive process.The influencing factors are determined from three aspects:insider attributes,common focus,and shared emotions.The definition of interestrelated community and fan community in this study intersects,and the research model has broad applicability;(2)Climb the 293566 interactive comments in the video series of"The Most Disturbing Hitman in History" of the uploader "old tomato" in bilibili.Firstly build the interactive network and simplify it into a core interactive network,compare and analyze the social networks,both have the characteristics of scale-free and non-small world.The node degree centrality,betweenness centrality and closeness centrality help to discover the members of interest-related community with high social capital and strong interaction tendency,and tiering members.Secondly,the structured topic model is used to cluster and analyze 7 comment topics,and the results show that the topics are affected by video updates and comment types,and further obtain the topic and emotional characteristics of members;(3)Empirical exploration of the generation mechanism of the core interactive network of the bilibili interest-related community.The core interaction network of interest-related community has the characteristics of reciprocity,convergence,expansion,transitivity and closure.Community members with medal,using emoji,and the number of works positively affect their comments,the number of attention negatively affects their comments,the number of fans positively affects reception of comments,while the presence of medal and the use of emoji negatively affect their reception of comments.Members with the same topic and emotion are more likely to comment on each other,while members with opposite emotion have a higher degree of interaction.This study starts with the research of interactive network,deeply explores the topic of comments,and enriches the theory and method of the research on the interest-related community of bilibili.Based on the conclusion of the empirical study,this study starts from the four elements of the interactive ritual chain,and provides practical ideas for the creation of interest-related content and community operation of bilibili.