Empirical insights into the interaction effects of groups at high risk of depression on online social platforms with NLP-based sentiment analysis

Abstract

With the proliferation of digital technology and the increasing prevalence of social media, some users at high risk of depression have opted to seek solace, acceptance, and assistance in online communities. However, the extant research is deficient in terms of the segmentation of groups, particularly subcultural groups. By analyzing the “Super Hashtags” and “Tree Hole” groups on Sina Weibo from January to March 2023 using a crawler and the ERNIE 3.0-Base model for sentiment analysis, the study uncovers distinct sentiment profiles and interaction patterns, revealing significant correlations between interaction metrics and sentiment levels. The findings indicate that while there are no significant differences in sentiment levels between the two communities, the “Tree Hole” community exhibits greater sentiment variability. Moreover, the study identifies that interaction behaviors are closely linked to sentiment states, emphasizing the importance of understanding the complex dynamics between online interactions and mental well-being. These insights contribute to the development of more effective support mechanisms within online platforms for individuals at risk of depression.