Research on risk assessment and optimization of GCN supply chain financial network based on M estimation

With rapid development of supply chain finance, risk assessment has become a key link to ensure financial security. Traditional methods make it difficult to assess risks in complex network environments. Therefore, this study proposes a graph convolutional network (GCN) model based on M estimation for risk assessment and optimization of supply chain financial networks. In the background, supply chain finance involves multi-party participation, and information asymmetry and complex relationship networks increase the difficulty of risk assessment. As a robust statistical method, M estimation can effectively deal with outliers and noise data. When combined with GCN, it can make full use of network structure information and improve the accuracy of risk assessment. Experimental results show that the proposed model is significantly superior to the traditional method in risk assessment performance. On the simulated supply chain financial network data set, the prediction accuracy rate of the model reaches 92.3%, which is 8.5 percentage points higher than that of the traditional logistic regression model. At the same time, the verification of the real data set also shows the model's superiority and the F1 score of risk assessment reaches 0.89, which is 0.12 higher than the benchmark model. In addition, the model also performs well in optimizing the risk control strategy. Adjusting the network structure and parameters minimises the risk, and the risk reduction rate reaches 15.7%. This study provides a new supply chain financial network risk assessment method and a theoretical basis and practical guidance for financial risk control and optimization.