Credit risk analysis using artificial intelligence: evidence from a leading South African banking institution (original) (raw)
Related papers
Evaluating Credit Risk Using Artificial Neural Networks
2011
In credit business, banks are interested in learning whether a prospective consumer will pay back their credit. The goal of this paper is to classify the credit risk which an applicant can be categorized as a good or bad consumer using artificial neural networks, to enable all parties to take remedial action. The Feed-forward back propagation neural network is used to classify a consumer into two classes depending on selected parameters. One of the classes is credit worthy and likely to repay its financial obligation. The other class which is not credit worthy and whose applications for credit will be rejected due to a high possibility of defaulting on its financial obligation. Two well known and available datasets have been used (German and Australian dataset) to test the proposed neural network. The results of applying the artificial neural networks methodology to classify credit risk based upon selected parameters show abilities of the network to learn the patterns. In German dat...
Artificial Intelligence-Enhanced Credit Risk Assessment on Commercial Bank of Ethiopia
Thesis, 2024
Accurately identifying credit risk is essential for the successful operation and growth of any business. By accurately predicting an applicant's loan status, businesses can better understand the drivers of credit risk and develop informed market strategies to promote business expansion. The goal of this research is to develop a machine learning model that can reliably predict the loan status (approved or rejected) of applicants to the Commercial Bank of Ethiopia (CBE). The study utilized a dataset of 32,285 applicants with 10 attributes. To evaluate the performance of different classifiers, the overall accuracy was used as the primary metric. Supervised machine learning techniques including Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and Deep Neural Network (DNN) were applied to the applicant loan status prediction task, based on their widespread use in prior literature. Feature importance and correlation matrix analysis were used for feature selection. Additionally, the SMOTE technique was employed to balance the dataset. The results show that the RF classifier achieved the best overall performance, with an accuracy of 93.62%, a recall of 94.72%,
International Journal of Applied Operational Research - An Open Access Journal, 2015
One of the most important issues always facing banks and financial institutes is the issue of credit risk or the possibility of failure in the fulfillment of obligations by applicants who are receiving credit facilities. The considerable number of banks’ delayed loan payments all around the world shows the importance of this issue and the necessary consideration of this topic. Accordingly, many efforts have been made for providing an efficient model for more accurate evaluation and classification of applicants receiving credit facilities for valid decision making about granting or not granting these facilities to them. Different statistic methods have been applied for this purpose, such as Discriminant Analysis, Probit Regression, Logistic Regression, Neural Network and so on. Among these methods, Neural Network has been considered mostly because of its high flexibility in recent years. In this research, many efforts have been made to examine the efficiency of Logistic Regression an...
Accurately identifying credit risk is essential for the successful operation and growth of any business. By accurately predicting an applicant's loan status, businesses can better understand the drivers of credit risk and develop informed market strategies to promote business expansion. The goal of this research is to develop a machine learning model that can reliably predict the loan status (approved or rejected) of applicants to the Commercial Bank of Ethiopia (CBE). The study utilized a dataset of 32,285 applicants with 10 attributes. To evaluate the performance of different classifiers, the overall accuracy was used as the primary metric. Supervised machine learning techniques including Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and Deep Neural Network (DNN) were applied to the applicant loan status prediction task, based on their widespread use in prior literature. Feature importance and correlation matrix analysis were used for feature selection. Additionally, the SMOTE technique was employed to balance the dataset. The results show that the RF classifier achieved the best overall performance, with an accuracy of 93.62%, a recall of 94.72%,
Neural nets versus conventional techniques in credit scoring in Egyptian banking
Expert Systems with Applications, 2008
Neural nets have become one of the most important tools using in credit scoring. Credit scoring is regarded as a core appraised tool of commercial banks during the last few decades. The purpose of this paper is to investigate the ability of neural nets, such as probabilistic neural nets and multi-layer feed-forward nets, and conventional techniques such as, discriminant analysis, probit analysis and logistic regression, in evaluating credit risk in Egyptian banks applying credit scoring models. The credit scoring task is performed on one bank's personal loans' data-set. The results so far revealed that the neural nets-models gave a better average correct classification rate than the other techniques. A one-way analysis of variance and other tests have been applied, demonstrating that there are some significant differences amongst the means of the correct classification rates, pertaining to different techniques.
Credit Risk of Bank Customers can be Predicted from Customer's Attribute using Neural Network
International Journal of Computer Applications, 2017
The aim of this paper is to present a model based on Multi-layer perceptron neural networks to recognize bad or good credit customers. Credit risk is one of the major problems in banking sector. Banks are faced with credit Risk while doing their tasks. Credit risk is the probability of non-repayment of bank loan granted to lenders. Decreasing Credit Risk, banks may perform better duties and responsibilities successfully for the economic growth of the country. This study will help for a banker to select a right borrower for investing bank fund and hereby may reduce non-performing loan. Artificial neural network is used for loan applicants' credit risk measurement and the calculations have been done by using SPSS and WEKA software. Number of samples was 101 and 12 variables were used to identify good customers from bad customers. The results showed that, History of borrower (Defaulter or non-defaulter), amount of loan, type of collateral security (physical assets or financial assets) and Value of collateral security had most important effect in identifying classification criteria of good and bad borrowers. The main contribution of this paper is specifying for credit rating of bank customers in Bangladesh's banking sector.
A Neural Network Approach for Credit Risk Evaluation
The Basel Committee on Banking Supervision proposes a capital adequacy framework that allows banks to calculate capital requirement for their banking books using internal assessments of key risk drivers. Hence the need for systems to assess credit risk. Among the new methods, artificial neural networks have shown promising results. In this work, we describe the case of a successful application of neural networks to credit risk assessment. We developed two neural network systems, one with a standard feedforward network, while the other with a special purpose architecture. The application is tested on real-world data, related to Italian small businesses. We show that neural networks can be very successful in learning and estimating the bonis/default tendency of a borrower, provided that careful data analysis, data pre-processing and training are performed.
Measuring Credit Risk of Bank Customers Using Artificial Neural Network
2013
Abstract In many studies, the relationship between development of financial markets and economic growth has been proved. Credit risk is one of problems which banks are faced with it while doing their tasks. Credit risk means the probability of non-repayment of bank financial facilities granted to investors. If the credit risk decreases, banks will be more successful in performing their duties and have greater effect on economic growth of the country.
Credit Risk Analysis Using Machine Learning Techniques
It can be easily observed that the general public is putting in more and more loan requests in the banking system recently, which can be regarded as a positive development for the banks, while at the same time presenting a considerable risk. Accurate risk management in the banking and finance sector is related to efficient and optimized use of the current resources, assessment of possible risks and taking timely precautions. It is of utmost importance for the banks to predict the problematic loans in terms of long-term stability. Giving credits to the applicants is one of the fundamental activities of the banks, however; the same activity brings significant risks. As part of their founding purpose, the banks do not avoid taking risks, and they choose to manage them. The banks should perform their risk management in the way to keep the damages resulting from the amount of loans they give to a minimum. Considering the above and in order to speed up the lending procedures in banks while making advantageous decisions, different algorithmic models and classifications, machine learning techniques such as artificial neural networks were started to be used lately, data mining being at the first place. In this study, the accuracy of the applicants' eligibility status for loans was determined by making use of several machine learning techniques. The open-access dataset from the German Credit Data UCI was employed. Based on the 1000 customers in this study's dataset, a 75,60% success rate was achieved in the XGBoost classifier, which has the best success rate among the studies conducted with the XGBoost classifier previously. In addition, the success rate is the highest among the other algorithms used in various studies made.
Artificial Neural Networks with Gradient Learning Algorithm for Credit Scoring[#115847]-98260.pdf
Recently, credit scoring problems have come into prominence depending on growing the number of applicants. As known from literature, the traditional techniques are not sufficient to model this kind of problems accurately. For this reason, the researchers are still struggling to develop the novel techniques and improve the current ones to achieve better solutions. In this paper, credit scoring problem is handled by artificial neural networks (ANNs) because they provide flexible modeling procedure and superior performances in the nonlinear environments. However, the researchers mostly overlook some important requirements such as model complexity, overfitting and selection of optimization algorithm during training of ANNs. This paper presents an efficient procedure that allows estimating more robust credit scoring models by means of the information criteria and the early stopping approach based on the cross-validation technique. In the application section, ANNs are trained by various gradient based algorithms over German credit scoring data, and then their classification performances are compared with each other and logistic regression. According to results, the performance of ANNs is better than logistic regression.