商业银行非现场监管的模式识别方法  

Pattern recognition method of off-site commercial banking regulation

作　　者：熊熊[1] 张维[1] 任达[1] 任昆[1] 安瑛晖[1] 

机构地区：[1]天津大学管理学院,天津300072

出　　处：《系统工程学报》2004年第5期532-537,共6页Journal of Systems Engineering

基　　金：国家自然科学基金95重大资助项目课题(79790130);天津市自然科学基金资助项目(023600411);管理学院青年科研基金资助项目.

摘　　要：商业银行非现场监管在整个商业银行监管体系中占有中枢神经的地位.文章将商业银行非现场监管作为模式识别分类问题进行研究,提出了通过神经网络的建模,构造基于马哈拉诺比斯距离的联合判别模型的方法.并以实际数据为基础验证了模型的有效性,进行了商业银行分类识别.The off-site banking regulation system plays a key role in the entire banking regulation system. In this paper, we regard the off-site regulation as a kind of pattern recognition problem, and a new model based on neural network and Mahalanobiz distance discriminant analysis is presented. Using this model, we classified commercial bank from the practical data. In the model, the input vector is regulation indexes. The model training is carried on through BP arithmetic and Mahalanobiz distance discriminant analysis in order to reach a certain accurate rate. A new sample is the classes corresponding to neuron whose output vector has the minimal distance with the connection weight vector of neuron. The effectiveness of the proposed approach is clearly demonstrated by the experiment results.

关 键 词：非现场监管 模式识别 商业银行 分类 BP神经网络 判别分析 

分 类 号：F830[经济管理—金融学]