A novel synthetical hierarchical community paradigm for social network division from the perspective of information ecosystem

It has received significant attention to identify different groups in online social networks. The obstruction of information flow has led to the emergence of social polarization, and extremism, resulting in the separation of online social networks. Related research has focused on horizontal community division and vertical leader differentiation but lacks the cross horizontal and vertical structures. Hence, we propose a horizontal and vertical binary structure of communities and hierarchies (HVBSCH) defined as “synthetical hierarchical communities (SHCs)" in online social networks and present an analytical framework for SHCs in the Weibo information ecosystem based on the integration of the information ecology theory and the structural hole theory. A modified sampling graph convolutional network algorithm was put forth to obtain sample labels of real-world communities, which was further used for community detection together with users' social and attribute features collected from the Weibo platform regarding the hot event “Wu Yanni's false start” during the Hangzhou Asian Games. The results indicate that the collaborative effects of celebrities and media generate large and stable communities, requiring only few intermediary levels of dissemination to spread influence. Structural hole spanners within communities trigger the formation of subgroups, facilitating the acquisition and transmission of information across hierarchies, thus positively impacting the formation of SHCs. This study contributes to expanding researchers' perspectives on structures of online social networks. The analytical framework demonstrates superiority in acquiring community labels and features of real-world network users. Also, structural hole complements the information ecology theory by quantifying the information ecological niche, thereby contributing to bridging divergences among users in online social networks.