Credit scoring using neural networks and SURE posterior probability calibration

Matthieu Garcin () and Samuel Stéphan ()
Additional contact information
Matthieu Garcin: ESILV - École Supérieure d'Ingénierie Léonard de Vinci
Samuel Stéphan: ESILV - École Supérieure d'Ingénierie Léonard de Vinci, SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne) - UP1 - Université Paris 1 Panthéon-Sorbonne

Working Papers from HAL

Abstract: In this article we compare the performances of a logistic regression and a feed forward neural network for credit scoring purposes. Our results show that the logistic regression gives quite good results on the dataset and the neural network can improve a little the performance. We also consider different sets of features in order to assess their importance in terms of prediction accuracy. We find that temporal features (i.e. repeated measures over time) can be an important source of information resulting in an increase in the overall model accuracy. Finally, we introduce a new technique for the calibration of predicted probabilities based on Stein's unbiased risk estimate (SURE). This calibration technique can be applied to very general calibration functions. In particular, we detail this method for the sigmoid function as well as for the Kumaraswamy function, which includes the identity as a particular case. We show that the SURE calibration technique is able to calibrate the predicted probabilities as well as the classical Platt method.

Keywords: Deep learning; credit scoring; calibration; SURE (search for similar items in EconPapers)
Date: 2023-10-20
New Economics Papers: this item is included in nep-big, nep-cmp, nep-isf and nep-rmg
Note: View the original document on HAL open archive server: https://hal.science/hal-03286760v2
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://hal.science/hal-03286760v2/document (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:hal:wpaper:hal-03286760

Access Statistics for this paper

More papers in Working Papers from HAL
Bibliographic data for series maintained by CCSD ().