Abstract:

[Objective] In the era of big data, the governance of major cyber events has become increasingly challenging. Constructing knowledge graphs and conducting analysis can effectively support intelligent decision-making. [Method] This paper employs an improved multi-agent collaboration framework, taking a trade event as the research subject. First, multi-source heterogeneous data was collected using Python, followed by the construction of a knowledge graph based on the enhanced multi-agent framework. Finally, the network structural characteristics of the graph are analyzed using complex network theory. [Results] The results demonstrate that this multi-agent knowledge graph construction framework can efficiently extract triples from unstructured data. The constructed knowledge graph exhibits the characteristics of “economy as the foundation, politics as the framework, and multi-dimensional extension.” Network analysis reveals that it has typical power-law distribution features. [Conclusion] This study effectively improves the efficiency of knowledge graph construction, alleviates the limitations of traditional cyber event research, and provides support for its structured and quantitative analysis.

Key words: public opinion governance, knowledge graph, intelligent agent, large language model