Search

This paper shows how to improve the measurement of credit scoring by means of factor clustering. The improved measurement applies, in particular, to small and medium enterprises (SMEs) involved in P2P lending. The approach explores the concept of familiarity which relies on the notion that the more familiar/similar things are, the closer they are in terms of functionality or hidden characteristics (latent factors that drive the observed data). The approach uses singular value decomposition to extract the factors underlying the observed financial performance ratios of SMEs. We then cluster the factors using the standard k-mean algorithm. This enables us to segment the heterogeneous population into clusters with more homogeneous characteristics. The result shows that clusters with relatively fewer number of SMEs produce a more parsimonious and interpretable credit scoring model with better default predictive performance. [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. [ABSTRACT FROM AUTHOR]

In the practice of credit-risk management, the models for receiver operating characteristic (ROC) curves are helpful in describing the shape of an ROC curve, estimating the discriminatory power of a scorecard, and generating ROC curves without underlying data. The primary purpose of this study is to review the ROC curve models proposed in the literature, primarily in biostatistics, and to fit them to actual credit-scoring ROC data in order to determine which models could be used in credit-risk-management practice. We list several theoretical models for an ROC curve and describe them in the credit-scoring context. The model list includes the binormal, bigamma, bibeta, bilogistic, power, and bifractal curves. The models are then tested against empirical credit-scoring ROC data from publicly available presentations and papers, as well as from European retail lending institutions. Except for the power curve, all the presented models fit the data quite well. However, based on the results and other favourable properties, it is suggested that the binormal curve is the preferred choice for modelling credit-scoring ROC curves. [ABSTRACT FROM AUTHOR]

This paper shows how to improve the measurement of credit scoring by means of factor clustering. The improved measurement applies, in particular, to small and medium enterprises (SMEs) involved in P2P lending. The approach explores the concept of familiarity which relies on the notion that the more familiar/similar things are, the closer they are in terms of functionality or hidden characteristics (latent factors that drive the observed data). The approach uses singular value decomposition to extract the factors underlying the observed financial performance ratios of SMEs. We then cluster the factors using the standard k-mean algorithm. This enables us to segment the heterogeneous population into clusters with more homogeneous characteristics. The result shows that clusters with relatively fewer number of SMEs produce a more parsimonious and interpretable credit scoring model with better default predictive performance.

Forecasting the creditworthiness of customers is a central issue of banking activity. This task requires the analysis of large datasets with many variables, for which machine learning algorithms and feature selection techniques are a crucial tool. Moreover, the percentages of "good" and "bad" customers are typically imbalanced such that over- and undersampling techniques should be employed. In the literature, most investigations tackle these three issues individually. Since there is little evidence about their joint performance, in this paper, we try to fill this gap. We use five machine learning classifiers, and each of them is combined with different feature selection techniques and various data-balancing approaches. According to the empirical analysis of a retail credit bank dataset, we find that the best combination is given by random forests, random forest recursive feature elimination and random oversampling. [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.

In the practice of credit-risk management, the models for receiver operating characteristic (ROC) curves are helpful in describing the shape of an ROC curve, estimating the discriminatory power of a scorecard, and generating ROC curves without underlying data. The primary purpose of this study is to review the ROC curve models proposed in the literature, primarily in biostatistics, and to fit them to actual credit-scoring ROC data in order to determine which models could be used in credit-risk-management practice. We list several theoretical models for an ROC curve and describe them in the credit-scoring context. The model list includes the binormal, bigamma, bibeta, bilogistic, power, and bifractal curves. The models are then tested against empirical credit-scoring ROC data from publicly available presentations and papers, as well as from European retail lending institutions. Except for the power curve, all the presented models fit the data quite well. However, based on the results and other favourable properties, it is suggested that the binormal curve is the preferred choice for modelling credit-scoring ROC curves.

Forecasting the creditworthiness of customers is a central issue of banking activity. This task requires the analysis of large datasets with many variables, for which machine learning algorithms and feature selection techniques are a crucial tool. Moreover, the percentages of “good” and “bad” customers are typically imbalanced such that over- and undersampling techniques should be employed. In the literature, most investigations tackle these three issues individually. Since there is little evidence about their joint performance, in this paper, we try to fill this gap. We use five machine learning classifiers, and each of them is combined with different feature selection techniques and various data-balancing approaches. According to the empirical analysis of a retail credit bank dataset, we find that the best combination is given by random forests, random forest recursive feature elimination and random oversampling.

The main aim of the research was to determine the key factors determining the level of credit risk of individual clients (clients in the form of natural persons, excluding companies) on the example of Polish cooperative banks according to the following features: transaction characteristics, socio-demographic characteristics of the customer, the customer’s financial situation, the customer’s history of cooperation with the cooperative bank where they applied for a loan, and the customer’s history of cooperation with other financial institutions. For the research gathered data from 1000 credit applications submitted by individual customers when applying for a credit in five different cooperative banks were used for the analyses. To assess the credit risk of retail clients we use logit regression models, and additionally, score cards were calculated. The results of the research indicate that among the factors with high predictive power there were the features characterizing the client’s history of cooperation with the cooperative bank, where they applied for a loan. It may mean that when assessing credit risk related to financing individual customers, cooperative banks due to their local character, have an advantage over other financial institutions.

In micro-lending markets, lack of recorded credit history is a significant impediment to assessing individual borrowers’ creditworthiness and therefore deciding fair interest rates. This research compares various machine learning algorithms on real micro-lending data to test their efficacy at classifying borrowers into various credit categories. We demonstrate that off-the-shelf multi-class classifiers such as random forest algorithms can perform this task very well, using readily available data about customers (such as age, occupation, and location). This presents inexpensive and reliable means to micro-lending institutions around the developing world with which to assess creditworthiness in the absence of credit history or central credit databases.

Predicting if a client is worth giving a loan—credit scoring—is one of the most essential and popular problems in banking. Predictive models for this goal are built on the assumption that there is a dependency between the client’s profile before the loan approval and their future behavior. However, circumstances that cause changes in the client’s behavior may not depend on their will and cannot be predicted by their profile. Such clients may be considered “noisy” as their eventual belonging to the defaulters class results rather from random factors than from some predictable rules. Excluding such clients from the dataset may be helpful in building more accurate predictive models. In this paper, we report on primary results on testing the hypothesis that a client can become a defaulter in two scenarios: intentionally and unintentionally. We verify our hypothesis applying data driven regularized classification using an autoencoder to client profiles. To model an intention as a hidden variable, we propose an especially designed regularizer for the autoencoder. The regularizer aims to obtain a representation of defaulters that includes a cluster of intentional defaulters and unintentional defaulters as outliers. The outliers were detected by our model and excluded from the dataset. This improved the credit scoring model and confirmed our hypothesis.

There is an increasing influence of machine learning in business applications, with many solutions already implemented and many more being explored. Since the global financial crisis, risk management in banks has gained more prominence, and there has been a constant focus around how risks are being detected, measured, reported and managed. Considerable research in academia and industry has focused on the developments in banking and risk management and the current and emerging challenges. This paper, through a review of the available literature seeks to analyse and evaluate machine-learning techniques that have been researched in the context of banking risk management, and to identify areas or problems in risk management that have been inadequately explored and are potential areas for further research. The review has shown that the application of machine learning in the management of banking risks such as credit risk, market risk, operational risk and liquidity risk has been explored; however, it doesn’t appear commensurate with the current industry level of focus on both risk management and machine learning. A large number of areas remain in bank risk management that could significantly benefit from the study of how machine learning can be applied to address specific problems.

The main aim of the research was to determine the key factors determining the level of credit risk of individual clients (clients in the form of natural persons, excluding companies) on the example of Polish cooperative banks according to the following features: transaction characteristics, socio-demographic characteristics of the customer, the customer's financial situation, the customer's history of cooperation with the cooperative bank where they applied for a loan, and the customer's history of cooperation with other financial institutions. For the research gathered data from 1000 credit applications submitted by individual customers when applying for a credit in five different cooperative banks were used for the analyses. To assess the credit risk of retail clients we use logit regression models, and additionally, score cards were calculated. The results of the research indicate that among the factors with high predictive power there were the features characterizing the client's history of cooperation with the cooperative bank, where they applied for a loan. It may mean that when assessing credit risk related to financing individual customers, cooperative banks due to their local character, have an advantage over other financial institutions. [ABSTRACT FROM AUTHOR]

Predicting if a client is worth giving a loan—credit scoring—is one of the most essential and popular problems in banking. Predictive models for this goal are built on the assumption that there is a dependency between the client's profile before the loan approval and their future behavior. However, circumstances that cause changes in the client's behavior may not depend on their will and cannot be predicted by their profile. Such clients may be considered "noisy" as their eventual belonging to the defaulters class results rather from random factors than from some predictable rules. Excluding such clients from the dataset may be helpful in building more accurate predictive models. In this paper, we report on primary results on testing the hypothesis that a client can become a defaulter in two scenarios: intentionally and unintentionally. We verify our hypothesis applying data driven regularized classification using an autoencoder to client profiles. To model an intention as a hidden variable, we propose an especially designed regularizer for the autoencoder. The regularizer aims to obtain a representation of defaulters that includes a cluster of intentional defaulters and unintentional defaulters as outliers. The outliers were detected by our model and excluded from the dataset. This improved the credit scoring model and confirmed our hypothesis. [ABSTRACT FROM AUTHOR]

In micro-lending markets, lack of recorded credit history is a significant impediment to assessing individual borrowers' creditworthiness and therefore deciding fair interest rates. This research compares various machine learning algorithms on real micro-lending data to test their efficacy at classifying borrowers into various credit categories. We demonstrate that off-the-shelf multi-class classifiers such as random forest algorithms can perform this task very well, using readily available data about customers (such as age, occupation, and location). This presents inexpensive and reliable means to micro-lending institutions around the developing world with which to assess creditworthiness in the absence of credit history or central credit databases. [ABSTRACT FROM AUTHOR]

There is an increasing influence of machine learning in business applications, with many solutions already implemented and many more being explored. Since the global financial crisis, risk management in banks has gained more prominence, and there has been a constant focus around how risks are being detected, measured, reported and managed. Considerable research in academia and industry has focused on the developments in banking and risk management and the current and emerging challenges. This paper, through a review of the available literature seeks to analyse and evaluate machine-learning techniques that have been researched in the context of banking risk management, and to identify areas or problems in risk management that have been inadequately explored and are potential areas for further research. The review has shown that the application of machine learning in the management of banking risks such as credit risk, market risk, operational risk and liquidity risk has been explored; however, it doesn't appear commensurate with the current industry level of focus on both risk management and machine learning. A large number of areas remain in bank risk management that could significantly benefit from the study of how machine learning can be applied to address specific problems. [ABSTRACT FROM AUTHOR]