The Multi-attribute Impact of Hyperlinks in Blogs: An Emotion-Centric Approach

Abstract As a digital social medium, blogs have transformed into arenas where individuals can express their perspectives, concepts, and feelings. This research delves into the spread of information in blog posts related to the Uyghur discourse, utilizing topic modeling to classify and prioritize pertinent content. The analysis reveals a significant emphasis on the Chinese Government’s suppression of Uyghur Muslims, as demonstrated by the hyperlinks embedded within the blog network. We mapped the discovered topics onto a network of hyperlinks, forming a topic-induced hyperlink network. Additionally, we conducted a toxicity analysis, which revealed a high level of toxicity in blog posts during that period. A morality assessment was also performed, revealing exceedingly high scores for all vices in the blog posts. Furthermore, we conducted an emotion analysis on the text and image data, which underscored a dominant sentiment of sadness pervading the blog posts and corresponding images. This reflects the severity of the Uyghur persecution. The research highlights the effectiveness of topic modeling, toxicity evaluation, and hyperlink analysis in shedding light on intricate social issues. Future studies could augment these results by expanding keyword scope and refining image recognition models within the context of Uyghur discourse.