Search

Credit scoring or predicting bankruptcy is among the most crucial techniques for identifying high-risk and low-risk credit situations. Accordingly, enhancing the accuracy of bankruptcy prediction methods decreases the risk of inappropriate financial decisions. Also, increasing the accuracy of credit scoring models brings significant benefits such as improved turnover, credit market growth, proper and efficient allocation of financial resources, and sustained improvement of the profits of banks, investors, funds, and governments. Various statistical classification methods have been developed in the literature with different features and characteristics for more accurate bankruptcy prediction. However, despite all appearance differences in statistical classification approaches, they all adhere to a common idea and concept in their training procedures. The basic operation logic in whole-developed statistical classification methods focuses on maximizing a continuous distance-based cost function to yield the highest performance. Despite it being a common and frequently used procedure for classification purposes, it is an unreasonable and inefficient manner to achieve maximum accuracy in a discrete classification field. In this paper, a new discrete direction-based Logistic Regression that is a common statistical classifier method for bankruptcy forecasting is proposed. In the proposed Logistic Regression, in contrast to all traditionally developed statistical classifiers, the compatibility of the cost function and the training procedure is considered. While it can be shown overall that the performance of the presented discrete direction-based classifier will not be inferior to its continuous counterpart, an evaluation of the suggested classifier is conducted to ascertain its superiority. For this purpose, three credit scoring datasets are considered to assess the classification rate of the presented classifier. Empirical outcomes demonstrate that, as pre-expected, in all cases, the model put forward can attain a superior performance compared to conventional alternatives. These findings clearly demonstrated the significant influence of the consistency between the cost function and the training process on the classification capability, a consideration absent in any of the traditional statistical classification procedures. Consequently, the presented Logistic Regression can be considered an efficient alternative for credit scoring purposes to achieve more accurate results. [ABSTRACT FROM AUTHOR]

In this paper, a credit scoring integration model incorporating BRNN is used to study the credit scoring problem in automobile finance. Aiming at the problems of existing credit scoring models constructed with shallow architecture and the unidirectional limitation of RNN itself, this paper introduces a BRNN model that superimposes RNN models in two directions. The potential relationship between each credit feature is mined through logistic regression, extreme gradient boosting tree, and bidirectional recurrent neural network algorithms, and the final prediction output is linked to the customer’s overall credit to improve the prediction accuracy. In this paper, we study the application of a credit scoring model based on the improved BRNN model for an auto finance company. Data preprocessing techniques and feature screening methods are used to improve the BRNN model and construct the credit scoring model for auto finance at Company A. The BRNN model is the basis for Company A’s credit scoring model. Based on the comparison with other models, it is concluded that the automobile finance credit scoring IBRNN model constructed based on the improved BRNN model in this paper has an accuracy of 89.6% in classifying the user finance data of Company A on different datasets, which is a significant improvement compared with the other five models.

This study implements the multivariable logistic regression to develop a credit scoring model based on tenants' characteristics. The credit history of tenant is not considered. Rental information of tenants was collected from a landlord company in Malaysia. Parameters of the multivariable logistic regression were estimated by using the penalized maximum likelihood estimation with ridge regression since separation in training data was detected. The initial factors considered that affect tenants' credit score were their gender, age, marital status, monthly income, household income, expense-to-income ratio, number of dependents, previous monthly rent, and number of months late payment. However, the marital status factor was then excluded from the logistic regression model due to its low significance to the model. Meanwhile, a tenant's credit scoring model was generated by calculating the tenant's probability of defaulting. The main factors of the tenant's credit score are the number of months late payment, the expense-to-income ratio, gender, previous monthly rent, and age. There is no underfitting or overfitting in the proposed credit scoring model which means the model's bias and variance are low. [ABSTRACT FROM AUTHOR]

Copyright of Recherches en Sciences de Gestion is the property of ISEOR and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Logistic regression is a very popular binary classification technique in many industries, particularly in the financial service industry. It has been used to build credit scorecards, estimate the probability of default or churn, identify the next best product in marketing, and many more applications. The machine learning literature has recently introduced several alternative techniques, such as deep learning neural networks, random forests, and factorisation machines. While neural networks and random forests form part of the practitioner's model-building toolkit, factorisation machines are seldom used. In this paper, we investigate the applicability of factorisation machines to some binary classification problems in banking. To stimulate the practical application of factorisation machines, we implement the fitting routines, based on logit loss and maximum likelihood, on commercially available software that is widely used by banks and other large financial services companies. Logit loss is usually used by the machine learning fraternity while maximum likelihood is popular in statistics. Depending on the coding of the target variable, we will show that these methods yield identical parameter estimates. Often, banks are confronted with predicting events that occur with low probability. To deal with this phenomenon, we introduce weights in the above-mentioned loss functions. The accuracy of our fitting algorithms is then studied by means of a simulation study and compared with logistic regression. The separation and prediction performance of factorisation machines are then compared to logistic regression and random forests by means of three case studies covering a recommender system, credit card fraud, and a credit scoring application. We conclude that logistic factorisation machines are worthy competitors of logistic regression in most applications, but with clear advantages in recommender systems applications where the number of predictors typically outnumbers the number of observations. [ABSTRACT FROM AUTHOR]

After reaching the completion point of the Heavily Indebted Poor Countries (HIPC) initiative and the subsequent debt relief, Côte d'Ivoire decided to take the necessary measures to avoid a new public debt crisis. Ten years later, the country is at a crossroads with a rapidly growing debt. Given this situation, it is important to question the effectiveness of the methods of financing the budget deficit in Côte d'Ivoire. This article aims to develop a credit scoring model in order to analyse the efficiency of budget deficit financing modes using a sample of 3222 loan lines from the database of the Department of Public Debt and Grants of Ivory Coast. The results show that loans from bilateral and commercial banks can be considered as "risky" financing. This mode of financing has a less important impact on the probability of default. On the other hand, the results obtained with the estimation of the Logit model show that the probability of default is strongly reduced when the government is financed by bondholders and other debt instruments and multilateral institutions. [ABSTRACT FROM AUTHOR]

Credit scoring is widely used by financial institutions for default prediction, however, a significant portion of online credit loan customers have inadequate or unverifiable credit histories, making it difficult for financial institutions to make effective credit decisions. Since the widespread use of smartphones and the popularity of mobile applications, it is worth investigating whether mobile application usage behaviors (App behaviors) of customers can effectively predict online loan defaults. This paper proposes a combined algorithm of CNN and LightGBM, and establishes credit scoring models with App behaviors to evaluate the default risk of online credit loans based on logistic regression, LightGBM, CNN and the combined algorithm, respectively. The experimental results suggest that App behaviors have an obvious effect on the default prediction of customers applying for online credit loans, and the combined model outperforms the other models in terms of the area under the curve (AUC). Furthermore, integrated credit scoring models are developed by combining App behaviors with traditional scoring features. A comparison of the integrated models and the traditional scoring model indicates that the integrated models have achieved a significant improvement in classification performance and App behaviors can be a powerful complement to the traditional credit scoring model. [ABSTRACT FROM AUTHOR]

Credit scoring has become an important issue because competition among financial institutions is intense and even a small improvement in predictive accuracy can result in significant savings. Financial institutions are looking for optimal strategies using credit scoring models. Therefore, credit scoring tools are extensively studied. As a result, various parametric statistical methods, non-parametric statistical tools and soft computing approaches have been developed to improve the accuracy of credit scoring models. In this paper, different approaches are used to classify customers into those who repay the loan and those who default on a loan. The purpose of this study is to investigate the performance of two credit scoring techniques, the logistic regression model estimated on categorized variables modified with the use of WOE (Weight of Evidence) transformation, and neural networks. We also combine multiple classifiers and test whether ensemble learning has better performance. To evaluate the feasibility and effectiveness of these methods, the analysis is performed on Lending Club data. In addition, we investigate Peer-to-peer lending, also called social lending. From the results, it can be concluded that the logistic regression model can provide better performance than neural networks. The proposed ensemble model (a combination of logistic regression and neural network by averaging the probabilities obtained from both models) has higher AUC, Gini coefficient and Kolmogorov-Smirnov statistics compared to other models. Therefore, we can conclude that the ensemble model allows to successfully reduce the potential risks of losses due to misclassification costs.

The old-fashioned logistic regression is still the most used method for credit scoring. Recent developments have evolved new instruments coming from the machine learning approach, including random forests. In this paper, we tested the efficiency of logistic regression and XGBoost methods for default forecasting on a sample of 35,535 cases from 7 different business sectors of Italian SMEs, on a set of 28 banking variables and 55 balance sheet ratios for verifying which approach is better supporting the lending decisions. With this aim, we developed an efficiency index for measuring each model's capability to correctly select good borrowers, balancing the different effects of refusing the loan to a good customer and lending to a defaulter. Also, we computed the balancing spread to quantify the different models' efficiency in terms of credit costs for the borrower firms. Results show that different sectors report different results. However, generally speaking, the two methods report similar capabilities, while the cutoff setting can make a substantial difference in the actual use of those models for lending decisions. [Display omitted] • Logistic regression and XGBoost methods performances for default forecasting are compared: on a sample of 35,535 cases from 7 different business sectors of Italian SMEs. • The analysis is performed no a set of 28 banking variables and 55 balance sheet ratios for verifying which approach is mostly able to support the lending decisions. • The expected gross income of the lending activity is computed based on each model outcome. • An efficiency measure is proposed, obtained as the relative position of the considered estimator from 0 % of the always wrong model to 100 % of the perfect forecasting model. • Different cutoff values are considered for both models, evidencing the subsequent different results. [ABSTRACT FROM AUTHOR]

Logistic regression is a very popular binary classification technique in many industries, particularly in the financial service industry. It has been used to build credit scorecards, estimate the probability of default or churn, identify the next best product in marketing, and many more applications. The machine learning literature has recently introduced several alternative techniques, such as deep learning neural networks, random forests, and factorisation machines. While neural networks and random forests form part of the practitioner’s model-building toolkit, factorisation machines are seldom used. In this paper, we investigate the applicability of factorisation machines to some binary classification problems in banking. To stimulate the practical application of factorisation machines, we implement the fitting routines, based on logit loss and maximum likelihood, on commercially available software that is widely used by banks and other large financial services companies. Logit loss is usually used by the machine learning fraternity while maximum likelihood is popular in statistics. Depending on the coding of the target variable, we will show that these methods yield identical parameter estimates. Often, banks are confronted with predicting events that occur with low probability. To deal with this phenomenon, we introduce weights in the above-mentioned loss functions. The accuracy of our fitting algorithms is then studied by means of a simulation study and compared with logistic regression. The separation and prediction performance of factorisation machines are then compared to logistic regression and random forests by means of three case studies covering a recommender system, credit card fraud, and a credit scoring application. We conclude that logistic factorisation machines are worthy competitors of logistic regression in most applications, but with clear advantages in recommender systems applications where the number of predictors typically outnumbers the number of observations.

Copyright of Gestão e Regionalidade is the property of Universidade Municipal de Sao Caetano do Sul and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The act of lending is based on trust in the borrower to honour the obligation of paying back the lender. Greater spreads on credit operations may help predict the expected recovery of the credit, based on the sufficiency and liquidity of the guarantee. This study aims to understand how predictive models can provide different estimations of expected recovery based on the same data sets. It classifies credit by the formulation of a rule that describes the values of a categorical variable according to some specified definition. It finds that a simple logistic regression model can easily be extended to a multiple logistic regression model by integrating more than one prediction variable, which indicates increasing difficulty in obtaining multiple observations with an increasing number of independent variables. It compares the efficiency of the logistic regression with that of a linear regression in predicting whether recovery is due in a credit operation, and, thus, identifies the best model for this purpose. [ABSTRACT FROM AUTHOR]

The article proposes a model of an effective credit management system for a credit institution. The goal is to determine the impact of selected loan management processes on the effectiveness of the entire credit management system. This is done through hypothesis testing, using the usual least squares method. Another problem is the assignment to individual clients of their real strategic importance in the credit portfolio of a credit institution in order to ensure the optimal allocation of financial resources. For this purpose, two different methods are used, namely: analysis of the client loan portfolio and the method of estimation using logistic regression. A model of an effective credit management system in a credit institution has been proposed. The factors affecting the credit management process and the effectiveness of the entire credit management system are identified. The strategic values of the credited clients for the credit institution are determined, which allows optimizing the distribution of the financial resources of the credit institution and the entire credit management system.

Currently, in the banking system, defaults in the repayment of loans have become one of the biggest problems, and banks and financial institutions have faced many problems such as the increase in the volume of outstanding receivables due to the lack of an appropriate system for allocating facilities. Considering the importance of credit risks, commercial banks used to apply judgment methods for determining those risks. However, the use of these methods was not efficient enough due to limited human abilities and, at the same time, various factors affecting credit risks in contrast to statistical methods as well as artificial intelligence methods. For this reason, this article measures the efficiency of logistic regression models and artificial neural networks in detection of bank customers’ credit status in the period of 2009-2013. The results indicated that the total accuracy rates of the Artificial Neural Network model and the Logistic Regression model were 87% and 77.2% respectively, and the error types I and II were reduced significantly in the neural network. According to the results, statistical models cannot be expected to properly evaluate the credit risk of customers with classical assumptions such as the linear relationship between variables. Therefore, application and integration of artificial intelligence techniques is strongly recommended in this regard.

The research is aimed at developing an algorithm for valuation by credit institution of the credit limit for a natural person-borrower, with regard to his socio-demographic characteristics, using the logistic regression tools. The improvement of the procedure for valuation of the solvency of an individual borrower is to integrate an empirical approach to the definition of solvency (based on the ratio of official income and expenditure of an individual) with the economic and mathematical methods, allowing to consider the available retrospective data of the financial institution about borrowers and to identify statistically significant regularities. The key role in the algorithm of estimation of credit limit is occupied by development of the logit-model for classification of borrowers for classes of «reliable» and «unreliable», the classification force of which will be to a large extent determinative as to the predictive quality of defining the amount of the credit limit. The publication presents an example of application of the logit-model for credit scoring, including quantitative and qualitative variables, with the help of which the socio-demographic characteristics of borrowers are described. The procedure of using the proposed algorithm is illustrated on the basis of such logit-model.

It is difficult for micro and small entrepreneurial firms (MSEFs) to access external financing from formal financial institutions because financial institutions cannot obtain sufficient and reliable credit information about MSEFs. With the development of the internet and data collection technologies, more and more data can be accessed from different sources, and logistic regression model often suffers from bad performance. In this paper, we propose a credit scoring model using composite MCP logistic regression and firstly apply the proposed method to predict the probability of default of MSEFs in China. The proposed method can carry out parameter estimation and automatic bi-level selection of variables simultaneously with respect to the grouping structure of variables. Empirical results of MSEFs with complex grouping structure data have shown that the proposed method outperforms forward stepwise logistic regression, MCP logistic regression and group MCP logistic regression. In addition, we find that gross salary, tax paid, bank and injury insurance information of MSEFs are the most important factors to predict the probability of default. [ABSTRACT FROM AUTHOR]

Copyright of Revista Gestão & Tecnologia is the property of Revista Gestao & Tecnologia and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Credit scoring is an important process for peer-to-peer (P2P) lending companies as it determines whether loan applicants are likely to default. The aim of most credit scoring models is to minimize the classification error rate, which implies that all classification errors bear the same cost; however, in reality, there is a significant cost-sensitive problem in credit scoring methods. Therefore, in this paper, a new cost-sensitive logistic regression credit scoring model based on a multi-objective optimization approach is proposed that has two objectives in the cost-sensitive logistic regression process. The cost-sensitive logistic regression parameters are solved using a multiple objective particle swarm optimization (MOPSO) algorithm. In the empirical analysis, the proposed model was applied to the credit scoring of a Chinese famous P2P company, from which it was found that compared with other common credit scoring models, the proposed model was able to effectively reduce type II error rates and total classification error costs, and improve the AUC, the F1 values (reconciliation average of Recall and Precision), and the G-means. The proposed model was compared with other multi-objective optimization algorithms to further demonstrate that MOPSO is the best approach for cost-sensitive logistic regression credit scoring models. First published online 27 November 2019

Esta investigación muestra la aplicación y desempeño de tres modelos para la clasificación de solicitantes de créditos: el modelo de análisis discriminante, el de regresión logística y el de redes neuronales; técnicas empleadas por las instituciones financieras en el cálculo del scoring de crédito. Los resultados obtenidos muestran un mejor desempeño del modelo de redes neuronales en comparación con el de regresión logística y análisis discriminante, logrando una tasa de aciertos en la clasificación del 86.9%. Para los tres modelos se emplearon catorce variables que informan sobre las características socioeconómicas del prestatario y sobre las características propias de la operación crediticia. En el ámbito de la gestión financiera, este resultado es importante dado que puede complementarse con el cálculo de la probabilidad de incumplimiento, con los montos expuestos en cada operación de crédito y con la tasa de recuperación de la entidad para establecer el valor de las pérdidas esperadas a nivel individual y a nivel del portafolio de créditos de la entidad. || This research shows the application and performance of three models for the classification of credit applicants: discriminant analysis, logistic regression and neural networks; techniques used by financial institutions for the calculation of credit scoring. The results show a better performance of the neural network model compared to logistic regression and discriminant analysis, achieving a success rate of 86.9\% in the classification. For the three models, fourteen variables were used to inform about applicant's socioeconomic characteristics and those of the credit operation. In the area of credit risk management, this result is relevant since it can be complemented by the calculation of default probability, the exposure at default and the recovery rate of the entity to establish the value of expected losses at both the individual level and the whole credit portfolio of the entity.

Credit scoring models have to fulfil two important aspects viz. good model performance and transparent explanation ability. Logistic Regression (LOGIT) has always been the standard credit scoring tool but the evolving economic trend has resulted in tricky data patterns that are difficult for LOGIT to quantify. Thus, Random Forest (RF) turns out to be receiving attention in this industry but having the black-box property as the main barrier. This note discusses the applicability of the RF in developing credit scoring model and the possible potential to be the main tool in the future. [ABSTRACT FROM AUTHOR]

Emerging markets contain the vast majority of the world's population. Despite the enormous number of inhabitants, these markets still lack a proper finance infrastructure. One of the main difficulties felt by customers is the access to loans. This limitation arises from the fact that most customers usually lack a verifiable credit history. As such, traditional banks are unable to provide loans. This paper proposes credit scoring modeling based on non-traditional-data, acquired from smartphones, for loan classification processes. We use Logistic Regression (LR) and Support Vector Machine (SVM) models which are the top linear models in traditional banking. Then we compared the transformation of the training datasets creating boolean indicators against the categorization using Weight of Evidence (WoE). Our models surpassed the performance of the manual loan application selection process, improving the approval rate and decreasing the overdue rate. Compared to the baseline, the loans approved by meeting the criteria of the SVM model presented a decreased overdue rate. At the same time, using the score generated by a SVM model we were able to grant more loans. This paper shows that credit scoring can be useful in emerging markets. The non-traditional data can be used to build robust algorithms that can identify good borrowers as in traditional banking. [ABSTRACT FROM AUTHOR]

Credit risk evaluation is getting importance due to competition and also the Basel requirements. This study is based on the evaluation of credit scoring models for retail bank loan customers of Askari Bank Limited in Pakistan. Credit risk models further brings efficiency in the evaluation of a loan application which ultimately reduces the overall risk of bank and also reduces the probability of default on loans. This study has used three credit scoring models of logistic regression, discriminant analysis and probit analysis for the evaluation of credit applicants. The ranking criteria for ranking among the various models is based on the average correct classification rate. Therefore, the ranking shows that probit analysis techniques has the highest average correct classification rate while Logistic regression technique has the lowest average correct classification rate. These statistical credit scoring approach further augments the judgmental approach of credit application evaluation procedure. Further studies can be carried out on an extended data sample of retail loans. Other important variables can be included in the model to further improve the validity of the credit evaluation procedure according to the regulatory requirements. The statistical procedures of decision trees and neural networks can also be applied in credit scoring procedures to further make it more authentic and reliable. [ABSTRACT FROM AUTHOR]

a lot of studies have been done about customer credit scoring, considering importance of the topic on credit institutions decision making. As an evolutionary computation method, Genetic algorithm is one of the methods used in this field. A variety of papers are published on comparing the performance of genetic algorithms with other scoring method but there is little information regard to fitness functions while these fitness functions play a vital role in overall performance of the model. To further investigation of the problem, three different fitness functions are proposed in the current paper and their performance is compared with other scoring methods including logistic regression and data envelopment analysis. The obtained results have shown that genetic algorithms quadratic function totally outperformed other methods based on accuracy, detection and sensitivity criteria.

One of the main tasks of the financial institutions is to give loan to the customers. Prediction and evaluation of the credit risks due to loan and consequently managing this risk is one of the greatest ongoing challenges for the financial institutions. The main aim of this work is to provide an optimized logistic regression model for credit scoring of real customers. Here the effects of increasing the customer’s credit classification from two (binary) to four (multinomial) distinct groups on the results of the logistic regression has been investigated. Identification of the most important parameters in prediction of the real customers’ credit scoring is the other important outcome of this work. The results of both binary and multinomial logistic regression show the relative importance of the education level and the age of the customer compared with other independent variables. The results of this work show that either increasing the number of classification types of the dependent variable, real customer’s credit, to four distinct groups has no sharp effect on the results of the optimized models or this conclusion can be due to improper distribution of the number of customers in different groups.

Abstract The aim of this research lies in the development of a credit-scoring model using logistic regression and multivariate discriminant analysis applied to 1500 individual loans from Moroccan MFIs. This model combines both behavioral and descriptive data related to the borrowers (age, activity, level of education, number of unpaid debts, number of loans, etc.) and others related to the institution (amount of credit, duration of credit, number of concluded loans per portfolio manager etc.). The results showed the importance of having a larger sample, a deep enough history on the behavior of the customer, and also more information about variables related to the client's activity and its performance to better predict the default. [ABSTRACT FROM AUTHOR]

Microcredit refers to small loans to borrowers who typically lack collateral, steady employment, or a verifiable credit history. It is designed not only for start-ups but also for individuals. The microcredit industry is experiencing fast growth in China. In contrast with traditional loans, microcredit typically lacks collateral, which makes credit scoring important. Due to the fast development of on-line microcredit platforms, there are various sources of data that could be used for credit evaluation. Among them, bank card transaction records play an important role. How to conduct credit scoring based on this type of data becomes a problem of importance. The key issue to be solved is feature construction: how to construct meaningful and useful features based on bank card transaction data. To this end, we propose here a so-called RFMS method. Here \R" stands for recency, \F" stands for frequency, and \M" stands for monetary value. Our method can be viewed as a natural extension of the classical RFM model in marketing research. However, we make a further extension by taking \S" (Standard Deviation) into consideration. The performance of the method is empirically tested on a data example from a Chinese microcredit company. [ABSTRACT FROM AUTHOR]

Copyright of Econometrics / Ekonometria is the property of Sciendo and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Behavioural scoring models are generally used to estimate the probability that a customer of a financial institution who owns a credit product will default on this product in a fixed time horizon. However, one single customer usually purchases many credit products from an institution while behavioural scoring models generally treat each of these products independently. In order to make credit risk management easier and more efficient, it is interesting to develop customer default scoring models. These models estimate the probability that a customer of a certain financial institution will have credit issues with at least one product in a fixed time horizon. In this study, three strategies to develop customer default scoring models are described. One of the strategies is regularly utilized by financial institutions and the other two will be proposed herein. The performance of these strategies is compared by means of an actual data bank supplied by a financial institution and a Monte Carlo simulation study. [ABSTRACT FROM AUTHOR]