Modeling and analysis of interest-tag-driven information cocoons formation mechanisms

The information cocoon phenomenon refers to the situation in social networks where, due to the combined effects of individual interest preferences and recommendation algorithms, information spreading becomes confined to specific interest groups, creating a closed information environment. To elucidate the formation mechanism of this phenomenon, this paper constructs a spreading dynamics model based on interest tags. We employ a unique encoding scheme to model user interest tags and introduce a tag-matching-based spreading mechanism to simulate interest-driven information spreading processes. Through simulation experiments, the study systematically analyzed the impact of different interest tags, spreading thresholds, and network types on the speed and coverage of information spreading. The results indicate that messages bearing prominent interest tags are more likely to exhibit local spreading within specific interest groups, while messages with weaker interest associations tend to exhibit greater spreading across groups. This mechanism-based discovery reveals the structural role of interest tags in the formation of information cocoons. Although the research is primarily based on simulation experiments, the findings align closely with the phenomenon of topic spreading on real-world social platforms. This provides a new theoretical perspective for understanding the evolutionary mechanisms of information cocoons and offers potential insights for optimizing recommendation systems and enhancing information diversity.