Analysis Of Active Fans Networks Based On Weibo Data Under Thematic Stimulation

The rapid growth of online social media in recent years has provided researchers and policy makers with a wealth of data documenting the daily activities,psychology and behaviour of users.Social networks,an important branch of complex networks,consist of nodes represented by social members and connected edges represented by social relationships.By analysing user data from social networks,it is possible to gain insight into user behaviour and preferences.In Sina Weibo,active fans refer to those who actively participate in discussions and interact with celebrities on Weibo,and they are an important group of users in Weibo because of their large number and wide age distribution.The study of active celebrity fans on Weibo can help us understand the social network under the emerging social media platform;it can help us understand the interaction behaviour of fans and the process of information dissemination in social media,and help policy makers control the online public opinion.In this thesis,we use the interaction data of celebrity active fans in Weibo,using celebrity active fans as nodes and interaction relationships as connected edges,to study the following three aspects from the perspective of complex networks in conjunction with Weibo themes: the interaction modes of celebrity active fans inspired by themes;the identification of core active fans in the process of active fan interaction;the influence of various themes on the amount of celebrity active fans and the growth patterns of active fans.In this thesis,Weibo topics were divided into six categories according to themes,namely endorsement,public welfare,music,sports,variety and national affairs,and nine representative celebrities with a large number of basic fans were selected from five different categories including actors,singers,online influencers,anchors and athletes.A total of 26,219,575 interactions between celebrities’ fans spanning from January to December 2021 were screened in Weibo,and a dynamically weighted active fans network was constructed for each celebrity with a half-hour interval for thematic stimulation.The topological characteristics of the active fans networks were first analysed from both static and dynamic perspectives.It was found that the active fans network is a heterogeneous scale-free sparse network,which means that the interaction between the star’s active fans under the theme-inspired network is dominated by the interaction between the ordinary fans(Non-hub nodes)in the star’s largest fan group(the largest connected group of the network under the theme-inspired network)and the high-impact active fans(Hub Nodes)and to a lesser extent between two or more active fans outside the largest fan base.Next,the core of the static active fan network was extracted by 6)-shell decomposition,and then the top 10 active fans were obtained by using the weighted modified PageRank algorithm for the core of the network.And the top 10 nodes of the same celebrity with different theme networks were counted and found that the core active fans of the celebrity were almost constant over the theme time span.Finally,the number of nodes in the dynamic active fans network changed at half-hourly intervals,and the relative growth rate and duration of growth of active fans were used to characterize the theme hotness and life cycle respectively,and the influence of the theme on the number of active fans of each celebrity was obtained with the help of these two indicators.The results of the study show that it is easier for celebrities to increase their traffic through the themes they are involved in by virtue of their identity or expertise,while it is harder for celebrities to increase their traffic through national affairs.It is also found that anchors are less likely to increase their fans’ attention through variety microblogs,but participation in public welfare microblogs can inspire more active fans.This thesis also uses the K-Shape time series clustering algorithm to obtain the growth patterns of the four types of active fans at the moment of explosive growth.After analyzing the proportion of the four patterns under the six categories of topics,it is found that the music theme only contains mode 3 single-peak decay and mode 4continuous decay,which leads to the short lifecycle of music-themed microblogs;the sum of mode 3 and mode 4 in the public welfare theme is the smallest,indicating that the fading rate of the heat of the public welfare theme is relatively slow.In addition,it was also found that fans were more active in promoting celebrities’ public welfare-themed microblogs.Through the analysis of social networks,this thesis investigates three aspects of celebrity active fans to help improve the understanding of Weibo social networks,reveal the characteristics of active fans’ preferences and the patterns of interaction formation in Weibo,help decisionmakers understand the inner mechanism of public opinion dissemination,provide guidance for various celebrities to increase the number of active fans and improve their own traffic,and also provide new perspectives and new ideas to explore the growth pattern and mobility of celebrity fans.