基于BP神经网络的商业银行流动性风险预警模型研究与应用

发布时间：2018-10-12 21:44

【摘要】：流动性是商业银行的生命线,流动性管理对于商业银行的稳健运营至关重要。商业银行在进行流动性管理时必然将流动性风险作为重点考察对象。只有将流动性风险控制在一定的范围内,同时保障商业银行较高的盈利性,商业银行的运营才能得益安全、稳定、持久进行。商业银行的流动性风险可以量化成一些具体的指标,在发生流动性危机时其相应的指标数据通常会有所反映,故而我们可以在流动性危机发生之前就对这些指标进行监控,实现流动性危机的预警。本文第一章阐述了我国商业银行流动性风险预警研究的背景及意义,同时梳理了国内外已有的关于商业银行流动性风险预警的文献。第二章对商业银行流动性风险预警的相关理论基础进行了阐述。第三章进行了商业银行流动性风险预警模型的需求分析和可行性分析。第四章在前文理论分析的基础上进行了商业银行流动性风险预警模型的设定。首先,对商业银行流动性风险预警机制做了统筹设计,对整体机制进行了构建;然后,在整体机制的基础上对具体模块进行了细化设计,其中包括构建商业银行流动性风险评价体系,以此确定输入模块和输出模块,构建BP神经网络预警模型,确定网络结构和激发函数。第五章采用样本银行的指标数据对BP神经网络预警模型进行了应用。首先采用样本银行数据对BP神经网络模型进行了训练,训练结果表明BP神经网络预警模型模拟效果较为理想,然后采用样本银行数据对BP神经网络模型的预警结果进行了检测,检测结果正确率为100%,说明该预警模型可以用于商业银行流动性风险预警。第六章进行结论总结,进一步阐述了本研究的解决的主要问题,并分析了研究中存在的不足,明确了下一步的研究方向。本文建立的BP神经网络预警模型在商业银行流动性风险预警方面构思较为完整,预测准确度较高,对我国现有商业银行流动性风险管理具有重要的现实意义。
[Abstract]:Liquidity is the lifeline of commercial banks. Liquidity management is very important to the steady operation of commercial banks. Commercial banks must focus on liquidity risk in liquidity management. Only by controlling the liquidity risk within a certain range and ensuring the commercial bank's higher profitability can the operation of the commercial bank benefit from safety stability and lasting. The liquidity risk of commercial banks can be quantified into some specific indicators, and when liquidity crisis occurs, the corresponding index data will usually be reflected, so we can monitor these indicators before the liquidity crisis occurs. Realize the warning of liquidity crisis. The first chapter expounds the background and significance of the research on liquidity risk early warning of commercial banks in China, and combs the literature on liquidity risk early warning of commercial banks at home and abroad. The second chapter expounds the theoretical basis of liquidity risk warning of commercial banks. The third chapter analyzes the demand and feasibility of the liquidity risk warning model of commercial banks. The fourth chapter sets the liquidity risk warning model of commercial banks on the basis of previous theoretical analysis. First of all, the liquidity risk warning mechanism of commercial banks is designed as a whole, and the overall mechanism is constructed. Then, on the basis of the overall mechanism, the detailed design of specific modules is carried out. It includes setting up the liquidity risk evaluation system of commercial banks, determining the input module and output module, constructing the early warning model of BP neural network, determining the network structure and excitation function. In the fifth chapter, the BP neural network early warning model is applied with the index data of the sample bank. First, the BP neural network model is trained with sample bank data. The result shows that the simulation effect of BP neural network early warning model is satisfactory, and then the early warning result of BP neural network model is detected by sample bank data. The accuracy of the detection results is 100, which indicates that the model can be used for the liquidity risk warning of commercial banks. In the sixth chapter, the conclusion is summarized, the main problems of this study are discussed, and the shortcomings of the research are analyzed, and the research direction of the next step is clarified. The BP neural network early-warning model established in this paper has a relatively complete conception and high prediction accuracy in the aspect of commercial bank liquidity risk warning. It has important practical significance for the liquidity risk management of commercial banks in China.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F832.33;TP183

【参考文献】