Abstract
<jats:p>With the rapid development of social e-commerce, user purchasing behavior is largely influenced by social relationships and group behavior, making herding effect a crucial factor in platform operation and consumer decision-making. This study constructs a quantitative model of herding effect in social e-commerce based on multi-agent modeling and social network analysis. Using user purchase data, social interaction relationships, and product popularity, the study calculates the group consistency index Gj and individual response coefficient Ri to systematically analyze user behavior. The results show significant differences in the group consistency index among different product categories. Electronic products (Gj=0.68) exhibit the strongest herding effect, while books and stationery (Gj=0.40) show relative independence. Optimizing intervention strategies through information prompts and recommendation algorithms can effectively regulate user behavior. Under the influence of these strategies, the purchase probability decreases by approximately 8% for highly sensitive groups, 5% for moderately sensitive groups, and 2% for low-sensitive groups. This study not only reveals the formation mechanism of herding effect in social e-commerce but also proposes quantifiable and operable intervention strategies, providing theoretical and practical basis for platforms to optimize user experience, improve sales structure balance, and maintain a healthy business ecosystem.</jats:p>