Comparison of different approaches using Random Forest for imbalanced credit data

Anna Matuszyk
Additional contact information
Anna Matuszyk: Warsaw School of Economics, Collegium of Management and Finance, Financial System Department

Bank i Kredyt, 2023, vol. 54, issue 4, 419-436

Abstract: Credit scoring models are extensively used in credit risk management of individual customers. These models are based on econometric methods using past data about customers, both defaulters and non- -defaulters. These models focus on the optimal separation between good and bad customers taking into account two types of errors that appear, namely: the False Positive (Type 1 error) and the False Negative (Type 2 error). The purpose of the project was to focus on the problem of unbalanced data. Different balancing methods have been applied to the data set obtained from the financial institution operating in the European market. Various levels of unbalance have been considered and different statistical assessment metrics have been compared.

Keywords: credit scoring models; unbalanced data; balancing technique; Random Forest; model performance (search for similar items in EconPapers)
JEL-codes: C01 C13 C52 (search for similar items in EconPapers)
Date: 2023
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://bankikredyt.nbp.pl/content/2023/04/bik_04_2023_04.pdf (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:nbp:nbpbik:v:54:y:2023:i:4:p:419-436

Access Statistics for this article

More articles in Bank i Kredyt from Narodowy Bank Polski Contact information at EDIRC.
Bibliographic data for series maintained by Wojciech Burjanek ().