基于IGA-BP神经网络和专家系统的我国银行流动性风险预警研究

以因子分析主成分方法选取预警指标;以免疫遗传算法对BP神经网络收敛速度较慢、易陷入局部极小等缺陷进行改进,形成IGA-BP神经网络.采用IGA-BP神经网络对银行的流动性风险进行预警,并提出以拥有强大知识库的专家系统对其输出的预警信号进行进一步核准和验证,确保预警信号的有效性,构建我国商业银行流动性风险预警机制.

A Study on Liquidity Risk Warning of Domestic Commercial Bank Based on IGA-BP Neural Network and Expert System

This paper uses principal component factor analysis to select early warning indicators.BP neural network has some defects,such as slowly converging speed and easily trapped into local minima.Using Immune Genetic Algorithm to improve them, and forming IGA-BP neural network,the generalization ability of neural network is enhanced.Then,IGA-BP neural network is used in the banking liquidity risk early warning,and output warning signals.At last,the expert system with a strong knowledge base is established to further approval and certification for warning signals,and ensures the effectiveness of early warning signals.Domestic commercial bank liquidity risk early warning system is constructed.