Answers are grounded in this paper and cite their sources.
Risk-warning prediction for manufacturing enterprises is essential for financial risk management, credit allocation, and sustainable industrial development. However, risk prediction for manufacturing enterprises suffers from two key limitations: the inadequate predictive accuracy of single-model approaches and the low credibility associated with “black-box” decision-making. To address these issues, this paper proposes an Interpretable Ensemble Learning Model (IELM) for risk-warning prediction of Chinese manufacturing enterprises, using firms’ risk-warning status as the prediction target. The framework integrates structured financial indicators with non-financial features extracted from annual report texts and supply chain characteristics, and constructs an optimized stacking ensemble via Bayesian hyperparameter optimization. Model interpretability is implemented using SHAP, which identifies how different indicators contribute to an enterprise being classified as risk-warning at both the aggregate and firm-specific levels. Retrospective stratified five-fold evaluation within the observed 2015–2020 sample period demonstrates competitive performance on the one-year risk-warning classification task, while the two-year and three-year analyses illustrate the model’s predictive capability over longer horizons. The interpretability analysis further reveals how key indicators, including the proportion of short-term borrowings to total assets, corporate governance, and annual report text, contribute to enterprises being classified as risk-warning or non-risk-warning, providing interpretable evidence on the indicators underlying manufacturing enterprise risk-warning predictions.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: