A comprehensive decision support approach for credit scoring
| Date | 03 October 2019 |
| Pages | 280-290 |
| Published date | 03 October 2019 |
| DOI | https://doi.org/10.1108/IMDS-03-2019-0182 |
| Author | Cuicui Luo |
| Subject Matter | Information & knowledge management |
A comprehensive decision support
approach for credit scoring
Cuicui Luo
International School, University of Chinese Academy of Sciences,
Beijing, China and
Stockholm Business School, Stockholm University, Stockholm, Sweden
Abstract
Purpose –The purpose of this paper is to provide a comprehensive decision support approach in credit
risk assessment.
Design/methodology/approach –A comprehensive decision support approach is proposed for credit
scoring and prediction. The predictive performance of the new approach has been investigated by using data
including number and text.
Findings –The results demonstrate that the proposed approach achieves better and more stable
classification accuracy than the single classifiers in most cases. Meanwhile, the prediction accuracy of
individual classifiers is also improved by the proposed approach.
Originality/value –This study provides a comprehensive model for credit risk scoring and provides
valuable information to the existing literature on credit scoring by using artificial intelligence.
Keywords Machine learning, Business intelligence, Risk analytics, Credit risk scoring, Decision support
Paper type Research paper
1. Introduction
The recent financial crises resulted in catastrophic loses globally and credit risk
management has become a major focus of banks and financial institutions. For example, it is
very important for banks to assess creditworthiness of a borrower before they lend a loan.
Due to the importance of credit risk management, a variety of analytical techniques have
been developed for credit risk assessment and credit scoring. Even though credit-rating
agencies such as Standard and Poor (S&P) and Moody have provided rating services for
many years, they are accused of being too slow to adjust their ratings and of being biased in
favor of borrowers (Hau et al., 2013; Löffler, 2005). In order to reduce the firm’s reliance on
external ratings and improve reliability, many business intelligence tools have been utilized
in risk management.
Various machine-learning techniques have been successfully used to deal with big data
in corporate credit scoring and prediction. Recent studies have shown that artificial
intelligent (AI) techniques, such as artificial neural networks (ANNs) (Atiya, 2001) and
support vector machine (SVM) (Bellotti and Crook, 2009), are superior to that of statistical
techniques in dealing with credit-scoring problems (Saberi et al., 2013; Lessmann et al., 2015).
In recent years, there has been an increased use of ensemble classifiers to improve
the prediction accuracy in credit-scoring applications. Recent examples are neural
network ensembles, random forests (RF) (Nanni and Lumini, 2009; Mercadier and Lardy,
2019), boosting and bagging ensembles. The performance of ensemble of classifiers in
credit-scoring problems has been investigated by many researchers and several evidence
suggest that the ensemble classifiers perform better than single classifiers in terms of
prediction accuracy in credit classification. For example, Brown and Mues (2012) conduct an
empirical study to investigate several classification algorithms in imbalanced credit-scoring
problem and the results indicate that the RF and gradient-boosting classifiers perform
better than single classifiers in a credit scoring. Jones et al. (2015) compare the predictive
performance of various binary classifiers using a large data set of corporate credit-rating
changes covering 30 years (1983–2013). They find that both AdaBoost and RFs have much
Industrial Management & Data
Systems
Vol. 120 No. 2, 2020
pp. 280-290
© Emerald PublishingLimited
0263-5577
DOI 10.1108/IMDS-03-2019-0182
Received 28 March 2019
Revised 6 May 2019
19 August 2019
Accepted 28 August 2019
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
280
IMDS
120,2
To continue reading
Request your trial