Influence of Social Networks in the Digital Economy Using Graph Neural Networks

In response to the problems faced by traditional social network influence analysis methods such as complex network structure, limited data sources, and lack of flexibility in analysis models, this article used graph neural network technology to conduct more accurate social network influence analysis in the digital economy field. A social network graph was constructed from two aspects: data preparation and graph construction, and then the social relationship features of nodes were extracted. Then, Graph Convolutional Networks (GCNs) were used for training to learn the information transmission patterns between nodes. Finally, the influence score of the node was calculated to identify its users with significant influence. The results showed that the influence score of the model increased over time from 4, 8, 11, 15, 21, and 26, indicating the reliability of the model's performance. The research results provide reliable support for social network analysis in the digital economy.