银行企业用户信用等级评价

发布时间：2018-09-06 19:06

【摘要】：随着国内经济持续繁荣发展,国内人民的消费能力越来越强,国内企业也越做越大。作为企业资金的中介载体,银行在整个国民经济的繁荣发展中起着越来越关键的作用。越来越多的企业从银行寻求贷款,并用所得贷款扩大生产规模,从而推动社会的发展。在所有的发展中的企业中,有的蓬勃向上,同样不可避免的也有黯然落幕。随着贷款企业越来越多,银行整体信贷的风险也越来越大,银行对企业的贷款额度也有考量。对企业的贷款额度的限定,决定了银行资金是否健康运转。当前银行对企业用户的贷款额度限制主要采用信用评级制度,信用等级越高,贷款额度越大,等级制度从根本上表明了企业的违约风险的大小。当前企业的信用等级的评定一般交给第三方组织,其评定过程不透明,极度依赖于第三方评定人员的从业经验,而企业的财务指标繁杂多变,不同规模的企业其财务规模也相差巨大,财务指标不尽相同,等级评定人员的一点失误就可能造成企业的等级的不同,从而对企业的发展产生巨大的影响。随着计算机技术的快速发展,许多产业开始摆脱原有的管理模式,在计算的管理下更加健康有序地发展。利用计算机来对企业进行信用评级也越来越可行。本文实现了一种基于BP神经网络的信用等级评价系统,首先利用银行收集的企业用户信息(定性指标、定量指标),对定量指标按时间顺序进行折旧预测并归一化处理,对定性指标进行量化分析,然后用BP神经网络对量化后的企业用户信息进行分类训练,最后用BP神经网络来对企业信用等级分类,从而得到企业的信用等级,并根据分类结果,得到信用等级评价的最小指标集合。本文的信用等级评价系统所用的数据来自企业实际数据,实际评级均为实际等级,本系统的信用评级与实际评级符合率大于90%,利用本信用评级系统,得到的最小指标集合只有87个指标,比原指标数少了20个。
[Abstract]:As the domestic economy continues to flourish, the consumption power of domestic people is becoming stronger and stronger, and domestic enterprises are becoming bigger and bigger. As the intermediary carrier of enterprise funds, banks play a more and more critical role in the prosperity and development of the whole national economy. More and more enterprises seek loans from banks and use the loans to expand the scale of production, thus promoting the development of society. In all the developing enterprises, some are thriving, and some are also inevitable and overshadowed. As the number of lending companies increases, the risk of overall bank credit is increasing, and the amount of loans granted by banks to enterprises is also taken into account. The limit on the amount of loans made by the company determines whether the bank funds are running healthily or not. At present, the credit rating system is mainly used to limit the loan quota to the enterprise users. The higher the credit grade, the larger the loan quota, which fundamentally indicates the size of the default risk of the enterprise. At present, the credit rating of an enterprise is generally given to a third party organization. The evaluation process is not transparent and depends heavily on the experience of the third party assessor, and the financial index of the enterprise is complex and changeable. The financial scale of enterprises with different scales is different, and the financial indexes are not the same. A bit of error of rating assessors may cause different grades of enterprises, which has a great impact on the development of enterprises. With the rapid development of computer technology, many industries begin to get rid of the original management mode and develop more healthily and orderly under the management of calculation. It is also more and more feasible to use computer to carry on credit rating to enterprises. In this paper, a credit rating evaluation system based on BP neural network is implemented. Firstly, the enterprise user information (qualitative index, quantitative index) collected by the bank is used to predict and normalize the depreciation of the quantitative index in chronological order. The qualitative index is analyzed quantitatively, then the enterprise user information is classified and trained by BP neural network. Finally, the enterprise credit grade is classified by BP neural network, and the credit grade of the enterprise is obtained according to the classification result. The minimum index set of credit rating is obtained. The data used in the credit rating evaluation system in this paper come from the actual data of the enterprise, and the actual ratings are all actual grades. The coincidence rate between the credit rating and the actual rating of this system is greater than 90, and the credit rating system is used to make use of the credit rating system. The minimum index set is only 87, which is 20 less than the original index.
【学位授予单位】：宁波大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F274;TP183

【相似文献】