基于神经网络评价模型的掌上贷款系统设计-计算机技术专业论文.docx

杭州电子科技大学硕士学位论文 杭州电子科技大学硕士学位论文 I I 摘 要 中小企业对我国的经济发展具有重要作用，然而由于我国现阶段金融市场以及社会信用 体系建设还不完善加之中小企业经营难以监管的特点，导致中小企业普遍存在融资困难的局 面，严重限制了中小企业的发展。 对于以上问题，本文首先对中小企业的经营特点以及银行贷款业务的流程进行了分析， 找出影响银行对中小企业贷款积极性的主要原因，针对这些影响因素，在基于银行现有贷款 业务流程的基础上，通过简化小微贷款业务流程并将其移植到移动客户端，来减小银行在小 微贷款中的成本投入和提高客户经理的工作效率。 同时针对中小企业贷款风险难控的问题，本文首先对现有预测模型建模方法进行了研究， 根据不同建模方法的使用环境，结合本文评价模型的输入是以定量数据为主，加之少量定性 数据的特点,选择采用 BP 神经网络来构建风险评价模型为银行贷款提供决策支持。然后在分 析了小微贷款企业的经营特点、影响小微贷款企业还款能力主要因素以及某银行贷款评价指 标的基础上，以工业企业为例制定了本文评价模型的评价指标体系，同时针对 BP 网络具有 的局部极小化问题，收敛速度慢等问题，文中又对 BP 网络算法做了部分改进，最后通过 Matlab 对评价模型进行了训练及测试，证明该模型具有一定的实用性。 关键词：小微贷款，贷款风险评价，手持终端，BP 神经网络 II II ABSTRACT The small and medium-sized enterprise is very important to the development of our country’s economic, However, due to the financial market and the social credit system is not perfect in China and the characteristics of small and medium-sized enterprise management, Small and medium-sized businesses widely has the problem of difficult loans which severely limited the development of some enterprise. To solve above problems, First, I analyse the characteristics of the small and medium-sized enterprises and the processes of bank loans to enterprises in order to find the the main causes for which the Bank do not want to provide loans to small and medium-sized enterprises. According to these factors, I simplify the process of bank loans to small and medium-sized enterprises and transplant the loan business to the mobile phones in order to reduce the cost of small loans in the investment and improve the working efficiency of the customer manager. Second, in order to solve the problem of micro-finance risk is difficult to control, I study the method to build the prediction models. Depending on the conditions of different method is used, I choose to use BP neural network to build the risk assessment model which will be used to provide help for bank lending decisions because the input data in this paper contains a large number of quantitative data and a small amount of qualitativ