Leveraging random forest in micro‐enterprises credit risk modelling for accuracy and interpretability

Abstract

This paper applies the Random Forest (RF) method for the robust modelling of credit default prediction. This technique has been proven as an efficient classifier and can provide better interpretability in comparison to other classifiers. Using Chines micro-enterprise credit data set, this study emphasizes the multidimensional analysis of credit risk, such as the whole sample, subsample, and the incremental effect of the group of predictors. To that end, relative variable importance (RVIs) has been presented for all predictors according to the contribution in the prediction accuracy so that to ensure interpretability of the model. The empirical findings confirm that RF technique is reliable and efficient across all of the criteria used in this study. In addition, the examined experimental analysis indicates that non-traditional variables have a significant effect on the classification accuracy. Thus, this paper recommends some alternative predictors like the legal representative's basic information, internal non-financial factors, along with traditional financial variables for sustainable model development. The performance is compared from the perspective of five different performance measures. This modelling algorithm can be used by different financial markets participants to measure systematically credit default prediction of individual and institutional customers.

Credit Risk Assessment of e-Commerce Supply Chain Finance of SMEs Based on Dynamic Reward and Punishment Perspective

“…ere should be a reasonable and comprehensive index system and appropriate methods for credit evaluation. Previous scholars mainly conducted the following studies in terms of the index system building and credit evaluation methods selecting: In terms of index system building, some scholars built is suitable for the supply chain under the background of the financing of small microenterprise credit evaluation index system [2][3][4][5][6][7][8]. Different from the traditional credit rating, the establishment of a credit risk evaluation system from different perspectives in the supply chain financial environment provides new ideas for the credit rating of SMEs [9].…”

Modeling credit risk with a multi‐stage hybrid model: An alternative statistical approach

“…Motivated by the above findings, this study employs LR, TreeNet ® , and RF to develop multi-stage hybrid models for credit default prediction. We consider TreeNet ® among the other methods of boosting because some recent studies mentioned that TreeNet ® , the advanced version of GB, outperforms other classifiers and its classification accuracy is better than other boosting studies (Cheng et al, 2018;Jiang & Jones, 2018;Jones, 2017;Uddin, Chi, Al Janabi, et al, 2020). Additional improvements such as XG boost are out of our scope for the same reason.…”

“…More recently, TreeNet ® has been effectively applied to credit default prediction by Uddin, Chi, Al Janabi, et al (2020), corporate distress prediction by Jiang and Jones (2018), bankruptcy prediction by Cheng et al (2018), and private company failure prediction by Jones and Wang (2019). Recent findings have confirmed the superiority of GB and RF approaches over frequently used AI methods such as, NN and SVM (Cheng et al, 2018;Jiang & Jones, 2018;Jones, 2017;Jones et al, 2015Jones et al, , 2017Uddin, Chi, Al Janabi, et al, 2020), but no evidence exists of hybrid or ensemble models that can be used to improve the prediction accuracy.…”

“…More recently, TreeNet ® has been effectively applied to credit default prediction by Uddin, Chi, Al Janabi, et al (2020), corporate distress prediction by Jiang and Jones (2018), bankruptcy prediction by Cheng et al (2018), and private company failure prediction by Jones and Wang (2019). Recent findings have confirmed the superiority of GB and RF approaches over frequently used AI methods such as, NN and SVM (Cheng et al, 2018;Jiang & Jones, 2018;Jones, 2017;Jones et al, 2015Jones et al, , 2017Uddin, Chi, Al Janabi, et al, 2020), but no evidence exists of hybrid or ensemble models that can be used to improve the prediction accuracy. However, the existing literature (in Table 2) reports on the deficiencies of single models and the superior capacity of hybrid classifiers compared to individual models (see, for example, Arifovic & Gencay, 2001;Blanco et al, 2001;Caigny et al, 2018;Chen & Li, 2010;Chi et al, 2019;Hamadani et al, 2013;Liang et al, 2015;Li et al, 2016;Oreski et al, 2012;Oreski & Oreski, 2014).…”

Credit Evaluation of SMEs Based on GBDT-CNN-LR Hybrid Integrated Model

“…e empirical results show that the probabilistic neural network (PNN) has the lowest classification error rate. Uddin et al [15] applied the random forest (RF) method to the robust modeling of credit default prediction, which has been proven as an efficient classifier than others. Wang et al [16] selected appropriate indicators and used an improved SVM model for analysis to be able to detect the credit risk of SMEs.…”