基于动态人工神经网络的投资银行风险预警  

Investment Banks Risk Early Warning Based on BP Neural Network

作　　者：杨丽敏[1] 郗英[1] 李淼[1] 

机构地区：[1]西北工业大学,陕西西安710072

出　　处：《安徽工业大学学报（自然科学版）》2006年第1期96-100,共5页Journal of Anhui University of Technology（Natural Science）

摘　　要：投资银行是金融体系的重要组成部分,从可量化监测的角度考察,选取了投资银行风险监测指标中包括市场风险、信用风险、流动性风险、资本风险在内的17个具体监测指标,构建了投资银行风险预警指标评价体系。运用基于数值优化的方法即L-M算法构建了基于前馈神经网络的投资银行风险预警模型。用训练好的BP神经网络模型,对检验样本进行了预测判别,结果显示出神经网络模型对我国投资银行风险具有较强的预测能力。Investment Banks is an important part of the whole financial system. Grasps 17 analysis indexes which include market risk, loans risk, liquidity risk and capacity risk, and a new analysis indexes early warning system is put forward. Also, ajudging method of artificial neural network which is based on L-M algorithm is given in this paper. The application shows that the analysis indexes system and BP artificial neural network can evaluate the risk condition of investment bank effectively.

关 键 词：投资银行 风险预警 BP神经网络 L-M算法 

分 类 号：F062.5[经济管理—政治经济学]