Linear and non-linear credit scoring by combining logistic regression and support vector machines

Gestel, Tony Van; Baesens, Bart; Dijcke, Peter Van; Suykens, Johan A. K.; Garcia, João

doi:10.21314/jcr.2005.025

Cited by 54 publications

(26 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most comprehensive are the empirical studies by Escott, Glormann, and Kocagil (2001), Falkenstein, Boral, and Carty (2000), and Sobehart, Keenan, and Stein (2000), who document nonlinear relationships in relation to finance structure (see also Estrella, Park, & Peristiani, 2000), profitability (see also Estrella et al, 2000;van Gestel et al, 2005), liquidity (see also Serrano-Cinca, 1997), turnover, growth (see also Hayden, 2011), and company size (see also Altman, Sabato, & Wilson, 2010). The most comprehensive are the empirical studies by Escott, Glormann, and Kocagil (2001), Falkenstein, Boral, and Carty (2000), and Sobehart, Keenan, and Stein (2000), who document nonlinear relationships in relation to finance structure (see also Estrella, Park, & Peristiani, 2000), profitability (see also Estrella et al, 2000;van Gestel et al, 2005), liquidity (see also Serrano-Cinca, 1997), turnover, growth (see also Hayden, 2011), and company size (see also Altman, Sabato, & Wilson, 2010).…”

Section: Literature Reviewmentioning

confidence: 99%

The total cost of misclassification in credit scoring: A comparison of generalized linear models and generalized additive models

Lohmann

Ohliger²

2018

Journal of Forecasting

View full text Add to dashboard Cite

This study examines whether the evaluation of a bankruptcy prediction model should take into account the total cost of misclassification. For this purpose, we introduce and apply a validity measure in credit scoring that is based on the total cost of misclassification. Specifically, we use comprehensive data from the annual financial statements of a sample of German companies and analyze the total cost of misclassification by comparing a generalized linear model and a generalized additive model with regard to their ability to predict a company's probability of default. On the basis of these data, the validity measure we introduce shows that, compared to generalized linear models, generalized additive models can reduce substantially the extent of misclassification and the total cost that this entails. The validity measure we introduce is informative and justifies the argument that generalized additive models should be preferred, although such models are more complex than generalized linear models. We conclude that to balance a model's validity and complexity, it is necessary to take into account the total cost of misclassification.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

The total cost of misclassification in credit scoring: A comparison of generalized linear models and generalized additive models

Lohmann

Ohliger²

2018

Journal of Forecasting

View full text Add to dashboard Cite

show abstract

“…7 The estimation of score functions and their validation are based on balance sheets of solvent and insolvent companies, whereas a company is classified as insolvent if it is the subject of failure judicial proceeding. The study is conducted over a long period, in order to construct durable scores that are resistant, as far as possible, to cyclical fluctuations.…”

Section: An Empirical Svm Model For Solvency Analysismentioning

confidence: 99%

Support Vector Machines (SVM) as a Technique for Solvency Analysis

2008

View full text Add to dashboard Cite

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractThis paper introduces a statistical technique, Support Vector Machines (SVM), which is considered by the Deutsche Bundesbank as an alternative for company rating. A special attention is paid to the features of the SVM which provide a higher accuracy of company classification into solvent and insolvent. The advantages and disadvantages of the method are discussed. The comparison of the SVM with more traditional approaches such as logistic regression (Logit) and discriminant analysis (DA) is made on the Deutsche Bundesbank data of annual income statements and balance sheets of German companies. The out-of-sample accuracy tests confirm that the SVM outperforms both DA and Logit on bootstrapped samples.

show abstract

“…10. The approach, introduced by Van Gestel et al for credit scoring [60][61][62], constructs an ordinal logistic regression model in a first step, yielding latent variable z L . 10.…”

Section: Incremental Approachmentioning

confidence: 99%

Rule Extraction from Support Vector Machines

Diederich¹

2008

Studies in Computational Intelligence

View full text Add to dashboard Cite

Linear and non-linear credit scoring by combining logistic regression and support vector machines

Cited by 54 publications

References 43 publications

The total cost of misclassification in credit scoring: A comparison of generalized linear models and generalized additive models

The total cost of misclassification in credit scoring: A comparison of generalized linear models and generalized additive models

Support Vector Machines (SVM) as a Technique for Solvency Analysis

Rule Extraction from Support Vector Machines

Contact Info

Product

Resources

About