基于增强关系图卷积网络的数据违规转售检测方法OA

Illicit data resale in data trading scenarios exhibited strong concealment and was difficult to detect.An en-hanced relational graph convolutional network was proposed by optimizing message passing and feature aggregation with transaction contextual similarity and causal temporal order constraints,enabling effective representation of illicit re-sale behaviors under complex transaction relations.Based on this model,a detection method was developed to predict the existence of illicit resale behaviors in transaction topology graphs.A simulated data trading dataset containing anoma-lous resale samples was constructed,and comparative experiments were performed.The results indicate that the pro-posed method provides an effective solution for illicit data resale detection in data trading scenarios.