模糊神经网络在纳税信用评估中的应用研究

由于纳税评估过程中存在不精确、模糊以及冗余信息,传统评估模型多数采用经验法和比较法,缺乏科学性和公正性,评估结果正确率低。为了提高纳税信用等级评估的正确率,提出了一种采用模糊神经网络的纳税信用等级评估模型。首先利用模糊逻辑推理对纳税评估过程中的不精确、模糊的信息进行有效的处理,然后利用训练数据对神经网络模型进行训练学习,获得纳税评估指标和信用等级间的评估模型,最后通过利用测试集对模型进行验证,结果表明,模糊神经网络方法提高了纳税信用等级评估的正确率,为税收信用评估提供有效的依据。     

神经网络的企业信用风险评估应用研究

研究企业信用风险评估准确性问题,企业存在产品质量、不良贷款等信用风险问题,企业信用风险是多种因素的综合结果,存在着不确定、非线性、随机性等特点,无法建立确定数学评估模型。只能根据专家评估指标为依据。为了提高企业信用风险评估准确率,提出一种BP神经网络的企业信用风险评估方法。先采用层次分析法构建风险评估指标体系,再用专家系统对评估指标进行量化打分,最后采用BP神经网络对企业信用风险指标进行非线性学习,并对企业信用风险等级进行评估。实验结果表明,BP神经网络的企业信用风险评估模模型能显著提高评估准确率,并能够反映企业信用风险的随机性变化特点,使评估结果更加符合实际情况,为企业信用风险评估提供了参考。     

金融风险预警的MPSO-FNN模型构建与应用

提出了一种改进型粒子群算法,并结合神经网络与模糊逻辑系统建立金融风险预警模型。将模型应用于信贷风险预警研究,仿真实例的结果表明,该模型所获得的预测准确性更高,是处理金融风险这类复杂经济系统预警问题的一种有效方法。     

Constructing credit auditing and control & management model with data mining technique

The 2008 financial tsunami, hitting the globe across all types of industries, causing tides of bankruptcies and severe unemployment, had its epicenter at American subprime in the housing market. In fact, the US subprime storm was just a premonition, while the root cause of the financial tsunami lied in the oversupply of structured credit products. Credit card business, one of the structured credit products, which under an intensively competitive environment, have been released by many banks with high spread, high return, and easy-to-apply appeals to carter to consumers needs. In order to allure the customers, some banks even go to the extent as simplify the credit rating, which in turn has increased credit risk, causing high non-performing ratio, increased debt collection cost, and growing bad debt counts. Accordingly, credit risk auditing plays a vital role in the successful management of credit card business. In response to such needs, the present study aims to conduct analysis and investigation on the current status of the industry with CRISP-DM model. First, customers’ demographic data and payment-related statistics were analyzed to identify feature variables, which were then sorted out as demographic data, debt data, payment rating etc. Next, by utilizing artificial neural network of data mining technique, the study tries to predict customer’s regular pattern of consumption, payment and/or default and bad debt, and to develop a set of credit granting principle by employing the decision tree technique. Since data mining classification model has a greater power in discriminating credit card granting, it can thus be used to construct accurate credit variable rules and predictive model, to further improve credit checking effect and credit risk control. Using the credit auditing data of a certain bank as a case study, the study intends to verify that the model constructed by the researcher can effectively identify the potential key factors of its credit card granting rule, to minimize the cost loss of Model I and Model II credit business, and eventually enhance the stability and profitability of the bank’s credit card business.     

基于改进型BP神经网络的信用评估系统研究

通过研究企业信用评估中的模型问题，为企业经营活动和决策过程提供信息支持。介绍了几种常用的信用评估模型，通过分析它们在评估中的缺陷，提出基于改进型BP神经网络的信用评估模型。在建立指标体系和输出机制的基础上，讨论了基于信用评估模型的评估系统的设计与实现。对模型和评估系统的不足进行了分析，并提出了改进建议。     

一般模糊极大-极小神经网络的研究与应用

分析一般模糊极大-极小神经网络的基本原理,阐述模糊计算方法在分类中的准确性和高效性。将一般模糊极大-极小神经网络应用于企业资信评估中,实现模糊区间的输入,缩小企业评估指标定量化中的误差范围。资信评估结果表明,该算法能快速、有效地对企业进行分类,为资信评估提供了解决方案。     

A credit ranking model for a parafinancial company based on the ELECTRE-III method and a multiobjective evolutionary algorithm

Credit rating is an assessment performed by lenders or financial institutions to determine a person’s creditworthiness based on the proposed terms of the loan. Frequently, these institutions use rating models to obtain estimates for the probabilities of default for their clients (companies, organizations, government, and individuals) and to assess the risk of credit portfolios. Numerous statistical and data mining methods are used to develop such models. In this paper, the potential of a multicriteria decision-aiding approach is studied. As a first step, the proposed methodology models the problem as a multicriteria evaluation process with multiple and in some cases, conflicting dimensions, which are integrated to derive sound recommendation for DMs. The second step of the methodology involves building a multicriteria outranking model based on ELECTRE III method. An evolutionary algorithm is used to exploit the outranking model. The methodology is applied to a small-scale financial institution operating in the agricultural sector. We compare loan applications based on their attributes and the credit profile of the customer or credit applicant. Our methodology offers the flexibility of combining heterogeneous information together with the preferences of decision makers (DMs), generating both relative and fixed rules for selecting the best loan applications among new and existing customers, which is an improvement over traditional methods The results reveal that outranking models are well suited to credit rating, providing good ranking results and suitable understanding on the relative importance of the evaluation criteria.     

基于神经网络的高校学生信用模糊评价研究

本文为更有效地对高校学生信用进行综合评价，建立了高校学生信用评价多级指标体系，提出了一种基于神经网络的评价模型。模型将模糊理论引入神经网络，通过神经网络的学习训练来调整模糊数学中的相应参数，以有效地模拟模糊综合评价的过程，对信用评价的定性过程而言，具有一定的合理性。     

面向高维数据的个人信贷风险评估方法

随着电商平台分期付款方式和P2P信贷平台的不断推广,如何从海量的用户信贷数据中挖掘出潜在的用户模型并对未知用户进行信贷风险评估,以降低信贷业务的风险,已经成为研究的主流。针对现有方法无法高效处理高维度信贷数据的问题,使用一系列的数据预处理方法和基于Embedded思想的特征选择方法XGBFS（XGBoost Feature Selection）,以降低用户信贷数据维度并训练出XGBoost评估模型,最终实现用户信贷风险评估。实验表明,与现有的方法相比,该方法能够从高维的数据中选择出重要属性,并且分类器在精确率、召回率等方面具有较为突出的性能。     

Fuzzy ART研究及在银行信用风险评估中的应用

探讨了Fuzzy ART神经网络的聚类功能及其参数对网络的影响。提出了一种基于该聚类理论的银行信用风险评估聚类模型。采用ASP.NET+MS SQL Server 2000的B/S构架实现了银行信用风险评估系统。通过上市公司财务数据验证了聚类结果的有效性和合理性。     

Fuzzy trust evaluation and credibility development in multi-agent systems

E-commerce markets can increase their efficiency through the usage of intelligent agents which negotiate and execute contracts on behalf of their owners. The measurement and computation of trust to secure interactions between autonomous agents is crucial for the success of automated e-commerce markets. Building a knowledge sharing network among peer agents helps to overcome trust-related boundaries in an environment where least human intervention is desired. Nevertheless, a risk management model which allows individual customisation to meet the different security needs of agent–owners is vital.The calculation and measurement of trust in unsupervised virtual communities like multi-agent environments involves complex aspects such as credibility rating for opinions delivered by peer agents, or the assessment of past experiences with the peer node one wishes to interact with. The deployment of suitable algorithms and models imitating human reasoning can help to solve these problems.This paper proposes not only a customisable trust evaluation model based on fuzzy logic but also demonstrates the integration of post-interaction processes like business interaction reviews and credibility adjustment. Fuzzy logic provides a natural framework to deal with uncertainty and the tolerance of imprecise data inputs to fuzzy-based systems makes fuzzy reasoning especially attractive for the subjective tasks of trust evaluation, business-interaction review and credibility adjustment.     

Municipal credit rating modelling by neural networks

The paper presents the modelling possibilities of neural networks on a complex real-world problem, i.e. municipal credit rating modelling. First, current approaches in credit rating modelling are introduced. Second, previous studies on municipal credit rating modelling are analyzed. Based on this analysis, the model is designed to classify US municipalities (located in the State of Connecticut) into rating classes. The model includes data pre-processing, the selection process of input variables, and the design of various neural networks' structures for classification. The selection of input variables is realized using genetic algorithms. The input variables are extracted from financial statements and statistical reports in line with previous studies. These variables represent the inputs of neural networks, while the rating classes from Moody's rating agency stand for the outputs. In addition to exact rating classes, data are also labelled by four basic rating classes. As a result, the classification accuracies and the contributions of input variables are studied for the different number of classes. The results show that the rating classes assigned to bond issuers can be classified with a high accuracy rate using a limited subset of input variables.     

支持向量机在信用卡信用评估的应用

信用卡业务现在是银行很重要的资产业务,构建一个适用的个人信用评估模型十分重要。基于近年来在智能学习系统领域发展起来的新理论．引入小样本学习的通用学习算法——支持向量机（SVM）,建立了个人信用评估模型,通过与神经网络模型的比较．证实了该方法用于信用卡个人信用评估的有效性及优越性。     

A study of Taiwan's issuer credit rating systems using support vector machines

By providing credit risk information, credit rating systems benefit most participants in financial markets, including issuers, investors, market regulators and intermediaries. In this paper, we propose an automatic classification model for issuer credit ratings, a type of fundamental credit rating information, by applying the support vector machine (SVM) method. This is a novel classification algorithm that is famous for dealing with high dimension classifications. We also use three new variables: stock market information, financial support by the government, and financial support by major shareholders to enhance the effectiveness of the classification. Previous research has seldom considered these variables. The data period of the input variables used in this study covers three years, while most previous research has only considered one year. We compare our SVM model with the back propagation neural network (BP), a well-known credit rating classification method. Our experiment results show that the SVM classification model performs better than the BP model. The accuracy rate (84.62%) is also higher than previous research.     

基于Levenberg-Marquardt神经网络的企业资信评估方法研究

企业资信评估问题是一个复杂的非线性问题,而神经网络技术可实现非线性关系的隐式表达。文章提出将基于Levenberg-Marquardt算法的多层前馈型神经网络用于资信评估,并通过MATLAB软件及其神经网络工具对其进行仿真计算。实验结果表明,企业资信神经网络评估模型收敛速度快,准确率较高,具有一定的实用价值。     

基于模糊神经网络的信息安全风险评估模型

对信息系统进行有效的风险评估,选择有效的防范措施,主动防御信息威胁是解决信息系统安全问题的关键所在。将神经网络和模糊理论应用于信息安全的风险评估。首先针对信息安全风险评估的不确定性和复杂性,将神经网络理论应用到风险评估。其次,针对神经网络适合定量数据,对于定性指标的分析缺乏相应的处理能力,而风险因素的指标值具有很大的不易确定性等问题,采用模糊评价法对信息安全的风险因素的指标进行量化,对神经网络的输入进行模糊预处理,提出了基于模糊神经网络的风险评估方法。仿真结果表明：模糊神经网络经过训练,可以实时地估算风险因素的级别。     

基于SVM混合集成的信用风险评估模型

准确的信用风险评估可以降低金融机构的风险。为了进一步提高信用风险评估模型的预测准确率,将基于SVM的集成学习模型应用到信用风险评估问题中,提出了一种混合集成策略,称作RSA。RSA是随机子集模型和AdaBoost两种流行策略的合成,能提高组合成员分类器的多样性,从而提高集成学习模型的预测准确率。模型在两组公开信用数据集上进行了应用,实验结果表明基于RSA的SVM的集成学习模型可以作为信用风险评估的有效模型。     

SOM神经网络模型在商业银行信用风险评估中的应用研究

研究了神经网络技术在商业银行信用风险评估中的应用,结合主成分分析法和SOM人工神经网络,建立了商业银行信用风险评估的人工神经网络模型;实证结果表明,该模型具有较高的预测精度.     

基于蚁群算法的神经网络在企业资信评估中的应用

BP算法在资信评估中应用较为广泛，但有收敛速度慢、易于陷入局部极小点的缺点。提出一种新的企业资信评估模型，该模型将蚁群算法和神经网络结合起来，使其既具有神经网络的广泛映射能力，又有蚁群算法带来的高效率，全局收敛，分布式计算等特点。实验表明,基于蚁群算法的神经网络对企业资信评估有着良好的性能。     

基于遗传算法的BP神经网络在企业资信评估中的应用

提出了一种新的企业资信评估方法.通过把神经网络和遗传算法有机地结合起来,既克服了传统BP网络训练时间长、易陷入局部极值的缺点,又利用遗传算法提高了网络全局收敛的效率.该模型采用C#.NET SQL server 2000实现.实验结果表明,基于遗传算法的BP神经网络系统对企业资信评估有着良好的性能.