“用户气泡”:社交网络使用中用户的异质性生产——基于k-means聚类的实证分析

“User Bubble”:The Heterogeneity Production on Users in the Use of Social Networks——Empirical Analysis Based on K-means Clustering

社交网络的快速发展正深刻形塑人类社会的“数字化生存”,关于它使人们变得同质化、标准化还是多元化、异质化的问题存在着不少争议。以新浪微博用户及其内容生产为样本对象,通过K-means聚类的方法得到样本用户及其帖子在不同主题类别上的分布,并根据使用程度对样本用户进行“分箱化”处理。研究结果发现,微博使用层级越高,用户相对于整体平均状态、本层级平均状态、最大众化层级、使用度最低层分别都会越发异质化;相对于最异质化层级、使用度最高层则越相似;同时用户不仅越来越多元,也越来越“窄化”,陷入一种“用户气泡”中。部分条件下,用户异质化的趋势表现为近乎直线性的增长。研究结果具备较突出的显著性。

The rapid development of social networks is profoundly shaping the“digital survival”of human and society.There are still a lot of controversies about whether social networks make people homogeneous,standardized or diversified,heterogeneous.This study takes Sina Weibo users and their contents as sample objects,obtains the distribution of sample users and their posts in different topic categories through k-means clustering algorithm,and divides users into different levels according to the degree of Weibo usage.We find that the higher the level of Weibo usage,the more heterogeneous the users will be compared with the overall users,the level which users belong to,the most popular level,and the level which has the lowest usage respectively.Users become more and more similar with the most heterogeneous level and the highest usage level following the rise of the Weibo usage’s level.In addition,users are not only more and more diverse,but also more and more“narrowed”,as a result,users are trapped in their“user bubble”.Under some conditions,the trend of heterogeneity almost shows linear growth.In general,the results are statistically significant.