Fusing Dynamics Equation: A Social Opinions Prediction Algorithm with LLM-based Agents

In the context where social media emerges as a pivotal platform for social movements and shaping public opinion, accurately simulating and predicting the dynamics of user opinions is of significant importance. Such insights are vital for understanding social phenomena, informing policy decisions, and guiding public opinion. Unfortunately, traditional simulation approaches struggle. This study proposes a groundbreaking solution to address this issue: the Fusing Dynamics Equation-Large Language Model (FDE-LLM) algorithm, which uniquely blends opinion dynamics with an epidemic modeling framework. This innovative approach effectively aligns the actions and evolution of opinions in Large Language Models (LLMs) with the true complexities of the cyber landscape. The FDE-LLM categorizes users into two roles: opinion leaders and followers. The Cellular Automata (CA) model manages opinions guided by LLM role-playing. In contrast, opinion followers are integrated into a dynamic system that combines the CA model with the Susceptible-Infectious-Recovered (SIR) model. This innovative design significantly improves the accuracy of the simulation. Our experiments utilized four authentic datasets from Weibo. The result demonstrates that the FDE-LLM significantly outperforms traditional Agent-Based Modeling (ABM) algorithms and LLM-based algorithms. The result highlights its superior accuracy and interpretability, underscoring its potential to revolutionize the understanding of social media dynamics.