An Artificial Neural Network Approach for Credit Risk Management

The objective of the research is to analyze the ability of the artificial neural network model developed to forecast the credit risk of a panel of Italian manufacturing companies. In a theoretical point of view, this paper introduces a litera-ture review on the application of artificial intelligence systems for credit risk management. In an empirical point of view, this research compares the architecture of the artificial neural network model developed in this research to an-other one, built for a research conducted in 2004 with a similar panel of companies, showing the differences between the two neural network models.

Identification of the Credit Guarantee Network of Steel Trade Enterprises in China

The risk points in the credit guarantee network of steel trade enterprises were identified by using the network analysis method in this paper. Firstly, the formation and operation mechanism of steel trade credit guarantee network was analyzed.Secondly,a guarantee network was established to analyze the related network structure indexes based on the mutual guarantee data of 83 enterprises in a steel trade market. These indexes included centrality,honest broker,and structural hole. The results suggest that network analysis method can be used to find out the risk points of the guarantee network. Additionally,some recommendations are brought forth to reduce or prevent future crises.

P2P network lending in the credit risk rating of the individual

P2P lending network is person to person lending network, lnternet-based applications, individuals lending financial model to others through the network intermediary,＇ platform. Currently P2P lending network has developed rapidly, but the P2P network lending platform also are lacing increasing risks, the biggest risk is credit risk. This article from the credit rating perspective, comparative analysis of the existing credit rating methodology, Analysis to establish a relatively sound credit rating mechanisms, thus reducing credit risk.

A Hybrid Approach for Predicting Probability of Default in Peer-to-Peer (P2P) Lending Platforms Using Mixture-of-Experts Neural Network

Peer-to-peer (P2P) lending offers an alternative way to access credit. Unlike established lending institutions with proven credit risk management practices, P2P platforms rely on numerous independent variables to evaluate loan applicants’ creditworthiness. This study aims to estimate default probabilities using a mixture-of-experts neural network in P2P lending. The approach involves coupling unsupervised clustering to capture essential data properties with a classification algorithm based on the mixture-of-experts structure. This classic design enhances model capacity without significant computational overhead. The model was tested using P2P data from Lending Club, comparing it to other methods like Logistic Regression, AdaBoost, Gradient Boosting, Decision Tree, Support Vector Machine, and Random Forest. The hybrid model demonstrated superior performance, with a Mean Squared Error reduction of at least 25%.

Supply Chain Finance Credit Risk Evaluation Method Based on Self-Adaption Weight

Credit risk is the core issue of supply chain finance. In the supply chain, problems happened in different enterprises can influent the whole to different degrees through transferring, thus statuses of all enterprises and their different influences should be considered when evaluating the supply chain’s credit risk. We examine the characters of supply chain network and complex network, use the local growing complex network to simulate the real supply chain, use cluster analysis to classify the company into several levels;Introducing each level’s self-adaption weight formula according to the company’s quantity and degrees of this level and use the weight to improve the credit evaluation method. The research results indicate that complex network can be used to simulate the supply chain. The credit risk evaluation (CRE) of an enterprise level with bigger note degrees has a greater weight in the supply chain system’s CRE, thus has greater effect on the whole chain. Considering different influences of different enterprise levels can improve credit risk evaluation method’s sensitivity.

MRBCH: A Multi-Path Routing Protocol Based on Credible Cluster Heads for Wireless Sensor Networks

Wireless sensor networks are widely used for its flexibility, but they also suffer from problems like limited capacity, large node number and vulnerability to security threats. In this paper, we propose a multi-path routing protocol based on the credible cluster heads. The protocol chooses nodes with more energy remained as cluster heads at the cluster head choosing phase, and then authenticates them by the neighbor cluster heads. Using trust mechanisms it creates the credit value, and based on the credit value the multi-path cluster head routing can finally be found. The credit value is created and exchanged among the cluster heads only. Theoretical analysis combined with simulation results demonstrate that this protocol can save the resource, prolong the lifetime, and ensure the security and performance of the network.

Fuzzy-Neuro Model for Intelligent Credit Risk Management

This paper presents hybrid fuzzy logic and neural network algorithm to solve credit risk management problem. Credit risk is the risk of loss due to a debtor’s non-payment of a loan or other line of credit. A method of evaluating the credit worthiness of a customer is complex and non-linear due to the diverse combinations of risk involve. To address this problem a credit scoring method is proposed in this paper using hybrid fuzzy logic-neural network (HFNN) model. The model will be implemented, tested, and validated for individual auto loans using real life bank data. The neural network is used as the learner and the fuzzy logic is used as the implementer. The neural network will fine tune the fuzzy sets, remove redundant input variables, and extract fuzzy rules. The extracted fuzzy rules are evaluated to retain the best k number of rules that will give final and intelligent decisions. The experiment results show that the perform-ance of the proposed HFNN model is very accurate, robust, and reliable. Comparison of these results to other previous published works is also presented in this paper.

Evaluation and Empirical Analysis of the Individual Housing Mortgage Loan Credit Risk

The paper applles the methods of discriminant analy-sis,logistic regression and artificial neural networkto the credit risk evaluation of individual housingmortgage loan made by commercial banks.On thebase of the comparison among the results of thesethree methods,the article analyzes their effectivenessand concludes that artificial neural network methodshould be the best choice.

基于信誉评分的共识网络重组机制设计

针对共识算法对良性节点占比的依赖,设计了一种基于信誉评分的共识网络重组机制,使得网络能在遭受攻击或发生故障时迅速自我修复和重组,显著提升了抗攻击能力和服务连续性。首先基于信誉机制理论,开发了一个分布式信誉评分系统,实时评估节点信誉,有效处理潜在恶意节点。此外,提出了一种基于信誉评分的共识网络重组机制,通过识别和隔离恶意节点来增强网络的弹性和安全性。该机制设计了网络弹性重组算法,整合了冗余设置、故障诊断和服务节点重新配置功能,确保网络稳定运行和高可靠性。该研究对金融服务、云计算和物联网等广泛使用共识技术的高安全需求领域具有重要应用价值。

基于CMCP-LMCL的多分类深度神经网络及其应用

多分类问题涉及信用风险管理、股票走势预测等多个领域。深度神经网络(DNN)是常用于多分类预测的机器学习模型,然而输入特征维度较高且存在冗余信号时,将加重其可解释性不强和结构冗余等缺陷;同时,常用的Softmax损失也可能面临分类边界模糊导致预测效果不佳等问题。为此,本文针对多分类问题,提出一个新的深度神经网络CMCP-LMCL,利用CMCP变量选择方法压缩输入特征到第1隐藏层的权重。该方法融合权重的组结构,能够剔除无关特征以及不重要的连接;同时,对特征层之外的权重施加权重衰减L;2;惩罚,有利于改进过拟合问题。新方法的增强边缘余弦损失(LMCL)在Softmax基础上引入扩大参数和距离参数,增大分类决策边界的间隔以期提高分类预测性能。模拟分析表明,对比已有DNN和传统分类方法,无论特征以简单线性形式还是复杂非线性形式映射到因变量,本文所提出的方法均具有良好的特征选择性能和预测表现。基于信用贷款数据的实证分析表明,该方法能够有效选择风险指标并进行违约风险预警。

终身学习视域下学分银行新角色扩展探析

学分银行是构建终身教育体系的重要组成部分,具有全民性、权威性、安全性和灵活性等优势。但同时,它也因管理功能单一、数据采集困难、管理体制突破困难等问题而导致发展面临困境。扩展学分银行新角色,构建智力资源供需平台,形成“请求—响应”的“人人学、人人教”终身教育供给服务模式,基于社会知识网络的新型学分银行进一步形成“个体需求—多方响应—共同体学习”的新型终身学习模式。学分银行扩展新角色对加快发展面向每个人、适合每个人、更加开放灵活的终身教育体系具有重要意义。

经济政策不确定性与商业信用融资--基于商业信用网络位置视角的分析

以2000—2022年沪深A股上市公司数据为样本,研究经济政策不确定性对商业信用融资的影响及商业信用网络位置的调节效应。研究发现,面对经济政策不确定性时,企业会选择提升商业信用融资,企业的商业信用网络中心位置能够帮助企业获得更多性商业信用融资。进一步研究表明,面对经济政策不确定性时,商业信用网络中心位置的企业是通过提升企业的议价能力与系统性风险地位进而提升商业信用融资的。面对经济政策不确定性提升,商业信用网络中心位置会缓解企业对银行信用的依赖,但当企业自身能够获得银行信用青睐时,商业信用网络中心位置会进一步提升企业的商业信用。因此建议企业在经济政策不确定高时充分融入商业信用网络,以商业信用资源弥补企业的银行信用不足,同时监管当局需要合理引导优化金融资源配置,避免商业信用与银行信用资源的过度倾斜。

银企信贷复杂网络下的银行业系统性风险研究——基于高碳行业视角

碳减排压力下,高碳企业向低碳转型内生的“转型风险”可能成为引发下一场系统性金融风险的“绿天鹅”事件。基于2013-2022年14个高碳行业的银企贷款数据,首先采用“银企”二分网络法,以高碳企业预期违约率作为风险传染强度指标,纵向度量“银企”的风险传导路径。研究发现银企间网络规模逐年扩大,尾部风险突出集中于化工行业,且主要由股份制商业银行承受。其次运用“银银”映射网络法,以银行间共同风险敞口作为风险传染强度指标,横向度量“银银”的风险传导路径。研究发现城市商业银行与农村商业银行日渐成为系统重要性银行,风险承担愈加凸显;国有控股大型商业银行自身输出较大银行间尾部风险,其对风险的扩散或防御至关重要。最后基于复杂网络筛选出系统重要性银行与风险传染性银行,运用动态CoVaR法度量尾部溢出的系统性风险贡献强度,全面捕捉风险从高碳企业到金融机构间“自下而上的风险测度逻辑”,从而为监管部门构建“自上而下的风险监管逻辑”提供现实依据。在“双碳”战略下,应出台专门的信贷支持政策推动重点行业实施节能减排,立足“银企”信贷网络的全局开展差异化管理和全面风险评估,在逆周期缓冲资本监管框架的挂钩变量中引入高碳行业尾部风险的新视角。

面向企业信用风险评估的多视角异质图神经网络方法

企业信用风险评估是一个重要且具有挑战的问题.由于金融市场中存在大量的异质关联关系,使得异质图神经网络天然适合建模企业信用风险.然而,现有大部分研究不能充分捕捉到复杂金融网络中企业的综合信用风险.针对此问题,提出了一个面向企业信用风险评估的多视角异质图神经网络方法——CRGNN.该方法包含自身风险编码器以及传染风险编码器,其中自身风险编码器建模基于企业特征信息的自身风险,传染风险编码器由新提出的分层异质图Transformer网络和分层异质图特征注意力网络2个子模块组成.这2个模块分别挖掘基于企业不同邻居视角的传染风险和基于不同特征维度视角的传染风险.为了充分利用异质关系信息,2个模块都采用了分层机制.在企业破产预测数据集SMEsD和企业信用评估数据集ECAD上进行了大量的实验,AUC指标相比最优基线模型分别提高了3.98个百分点和3.47个百分点.

个体借贷关系网络特征与违约研究——来自互联网融资平台的证据

本研究应用复杂网络分析方法研究了互联网融资平台个体借贷关系网络特征与借款人违约问题.提出了融资平台个体借贷关系网络的构建方法,构建了互联网借贷平台人人贷个体借贷关系网络,刻画了该网络的整体特征以及违约借款人群体和不同信用等级借款人群体、纯投资者群体构成的子网络特征.揭示了借款人和投资者的行为选择规律.分析了借款人网络拓扑特征与借款信息的相关性,构建了借款人违约与借款人网络拓扑特征、借款信息的三个回归模型.研究发现:构造的网络拓扑特征对于个体借贷关系结构具有很强的解释力;借款人节点的网络拓扑特征与借款人的违约存在显著相关关系,且这种关系在不同属性借款人群体中也具有稳健性;研究结果能够用于各类互联网融资平台借款人信用风险评估及违约预测.

Predicting Credit Card Transaction Fraud Using Machine Learning Algorithms

Credit card fraud is a wide-ranging issue for financial institutions, involving theft and fraud committed using a payment card. In this paper, we explore the application of linear and nonlinear statistical modeling and machine learning models on real credit card transaction data. The models built are supervised fraud models that attempt to identify which transactions are most likely fraudulent. We discuss the processes of data exploration, data cleaning, variable creation, feature selection, model algorithms, and results. Five different supervised models are explored and compared including logistic regression, neural networks, random forest, boosted tree and support vector machines. The boosted tree model shows the best fraud detection result (FDR = 49.83%) for this particular data set. The resulting model can be utilized in a credit card fraud detection system. A similar model development process can be performed in related business domains such as insurance and telecommunications, to avoid or detect fraudulent activity.

TSN中具有实时感知的AVB路由与调度算法

为了保障网络实时性传输需求,提升调度成功率,提出了具有实时感知的AVB流路由与调度算法,模拟车载时间敏感网络(TSN)场景下,音视频桥(AVB)流传输情况,并分析算法对网络调度成功率的影响。试验结果表明:随着数据流数量增加,相较于非实时感知算法与部分实时感知算法,具有实时感知的AVB流路由与调度算法使网络调度成功率分别提升26%和11%,该算法可优化TSN网络对AVB流的带宽预留,实现对数据流路由和报文信息的实时感知。

基于Transformer编码器和残差网络的信贷违约预测模型

针对传统信贷违约预测模型对高维稀疏类别特征缺乏有效处理,性能受到人工特征工程影响较大的问题,提出一种基于Transformer编码器和残差网络的信贷违约预测模型(TE-ResNet)。该模型首先利用嵌入层对类别特征进行处理,将它们转化为低维度的稠密向量;然后将连续特征和嵌入后的类别特征连接,输入到堆叠的Transformer编码器中进行特征提取,捕捉输入特征之间的关系,得到有用信息的高层特征表示;最后使用结合了通道注意力机制的一维残差网络模型进行违约预测。在训练过程中,模型采用加权交叉熵损失函数,以解决信贷数据不平衡的问题。实验结果表明,与8种主流基准模型的最佳表现相比,TE-ResNet在LendingClub数据集、天池贷款数据集上的各项指标均有提升:AUC指标分别提升了0.58%和2.85%,F1-Score指标分别提升了0.85%和11.92%,G-mean指标分别提升了2.94%和16.19%。TE-ResNet能够提高信贷违约预测的性能,减少人工特征工程,实现端到端的学习。因此,TE-ResNet模型具有实际应用的潜力,并可为信贷业务提供更加精确和可靠的风险评估服务。

基于图卷积网络的个人信用评估研究

传统的机器学习模型无法表现出个人信用评估问题中贷款者之间的高维邻居关系,通过构造图数据,利用图卷积网络进行分类预测的方式可以考虑贷款者在多个信息维度的相互联系。首先,利用递归特征消除(Recursive Feature Elimination,RFE)进行特征选择,筛选出贡献度最大的特征集。其次,利用随机森林(Random Forest,RF)计算出筛选后特征的重要性权重,同时将特征划分为类别特征和数值特征,根据特征类型并结合特征权重计算贷款者之间的距离,从而得到邻接矩阵。最后,将构造的图数据输入图卷积网络(Graph Convolutional Network,GCN)进行训练并预测结果。基于公开的德国个人信用数据集,通过两种评价指标对比了该模型与4篇近年研究结果,以及通过4种评价指标对比了该模型与3种基准模型。最终结果显示该模型的预测结果均要优于其他模型,能够有效进行个人信用贷款评估问题研究。

基于网络关系的分类变量预测研究

传统的预测方法通常基于个体的协变量信息进行建模和预测,少有考虑个体间的网络结构信息。事实上,网络节点间的关联信息能够为节点的响应变量预测提供信息,为此本文提出网络标签传播算法。基于半监督学习框架,以邻接矩阵为节点相似性推断依据,通过节点间的连接信息和已知节点的响应变量信息,来推断未知节点的响应变量信息。该算法适用于响应变量为分类变量的不完整网络数据。在网络服从随机分块模型的设定下,本文证明了该算法能够一致地预测未知节点的响应变量。数值模拟和实证数据分析结果显示,该算法预测效果较好。