基于图卷积算法的信贷风险防控研究

Research on credit risk control based on graph convolution algorithm

随着大数据时代的到来,我国互联网金融行业迅猛发展,传统的复杂网络分析法已经难以满足大数据背景下的信贷风险防控,而图卷积算法(GCN)旨在通过聚合邻居节点的属性和连接信息,将复杂网络的拓扑结构投影到低维向量空间中,能够有效挖掘和保存信贷用户网络的深层信息。因此,在信贷风险防控研究中引入图卷积算法,有利于提升金融机构信贷风险预测的准确性,促进信贷业务市场的健康发展。

With the advent of the era of big data,china's internet finance has been developing rapidly.Traditional complex network analysis method is difficult to meet the requirements of large-scale credit risk control.Graph convolution algorithm(GCN)projects the complex network to a low dimensional space by aggregating attributes and connection information of neighbor nodes.It can effectively mine and preserve the deep information of user network.Therefore,using graph convolution algorithm for credit risk control is conducive to improve the accuracy of credit risk prediction of financial institutions and promote the healthy development of the credit market.