Original scientific paper

https://doi.org/10.17559/TV-20251219003221

A Three-Stage Progressive Credit Risk Assessment Framework for Supply Chain Finance Based on Three-Way Decision and Ensemble Learning

Jianjun Hu ; School of Big Data and Artificial Intelligence, Guangdong University of Finance & Economics, Guangzhou, 510320
Haifeng Tan ; School of Big Data and Artificial Intelligence, Guangdong University of Finance & Economics, Guangzhou, 510320
Su Hu ; School of Big Data and Artificial Intelligence, Guangdong University of Finance & Economics, Guangzhou, 510320 *
Ruisheng Zhong ; School of Big Data and Artificial Intelligence, Guangdong University of Finance & Economics, Guangzhou, 510320 *
Wenhua Gan ; School of Big Data and Artificial Intelligence, Guangdong University of Finance & Economics, Guangzhou, 510320

* Corresponding author.

Abstract

The credit risk assessment of supply chain finance faces challenges such as insufficient data information, imbalanced categories, and insufficient fine-grained risk classification. To address the above issues, this paper proposes a three-stage progressive supply chain finance credit risk assessment framework by combining the three-way decision theory with the Multi Type Cascade Forest ensemble learning model. The framework introduces a delayed decision mechanism in the first stage to screen deterministic and uncertain samples. In the second stage, targeted supplementary information is provided through data augmentation strategies to reevaluate uncertain samples. In the third stage, the risk samples are further subdivided into risk levels and classified into low, medium, and high-risk enterprises. Using small and medium-sized manufacturing enterprises as the dataset, the experimental results show that compared with the traditional single-stage binary classification framework, the proposed framework improves the recall rate and F1 score of the model, and has better identification ability for high-risk enterprises, especially high-risk enterprises. The results indicate that combining progressive classification and ensemble learning can provide a more reliable and explanatory systematic decision-making method for credit risk assessment in supply chain finance.

Keywords

credit risk assessment; decision support systems; ensemble learning; progressive classification; supply chain finance; three-way decision

Hrčak ID:

353095

URI

https://hrcak.srce.hr/353095

Publication date:

7.11.2026.

Visits: 0 *