基于GA-ANN的中国金融安全预警系统设计及实证分析

国家金融体系的安全运行关系到经济社会的稳定,建立有效的金融安全预警系统已成为各界十分关注的焦点.基于现有文献,在金融安全预警指标体系中补充影子银行相关指标,以保证高杠杆、高流动性风险的经济参数参与建模,使得金融安全预警指标体系更加完整;运用因子分析计算七个金融子系统及整体金融系统安全得分,基于遗传算法优化的人工神经网络(genetic algorithm-artificial neural network,GA-ANN)建立中国金融安全预警系统,观察金融系统运行是否平稳、金融安全得分是否出现剧烈波动或异常值,以此判断国家金融状况是否安全,并对2013年我国金融安全状况进行预测.其中,GA-ANN网络较径向基神经网络、反向传播神经网络和广义回归神经网络,具有更好的拟合精度.预测结果显示2013年下半年我国金融系统总体运行安全,但在影子银行、股市和保险子系统存在一定的不安全因素.研究成果为政策制定者和广大投资者对国家宏观金融安全预判提供了参考依据.

Design of China's financial security early warning system based on GA-ANN

Economic and social stability is related to the security of the national financial system operation, the establishment of early warning system of financial safety has become the focus of concern from all walks of life. This paper combs the domestic and foreign literature; supplements the relevant indexes of shadow banks into the financial security early warning index system, which guarantees that high leverage and high liquidity risk economic parameters are involved in modeling, as well as making the financial early warning index system more complete; calculates seven financial subsystem and overall financial system safety score based on the factor analysis; sets up China's financial security early warning system based on the artificial neural network optimized by genetic algorithm (GA-ANN). In this paper, through the observation of whether the financial system running is stable or not and whether the financial safety is drastic fluctuations and abnormal value or not, so as to judge whether the national financial is safety, and carries on the forecast to the financial security in China in 2013. Compared with Radial Basis Function neural network, Back Propagation neural network and Generalized Regression neural network, the GA-ANN has better fitting precision. The prediction results show the macroscopic financial system in China in the second half of 2013 is safe, but there are some unsafe factors referring to shadow bank, stock and insurance system. The results of this study provide a reference for policy makers and investors when predicting the national macro financial safety.