Bankruptcy Prediction via Hybrid Resampling and Stacking Ensemble Techniques with Explainable Artificial Intelligence (XAI)-Driven Analysis

基于混合重采样、堆叠集成技术与可解释人工智能（XAI）驱动分析的破产预测研究

Abstract: This study develops and evaluates a bankruptcy prediction framework that integrates consensus-based feature selection, hybrid resampling, stacking ensembles, and explainable artificial intelligence to improve minority-class detection in severely imbalanced financial data.

摘要： 本研究开发并评估了一个破产预测框架，该框架整合了基于共识的特征选择、混合重采样、堆叠集成以及可解释人工智能，旨在改善严重不平衡金融数据中少数类别的检测能力。

Using the Taiwanese Bankruptcy Prediction dataset from the UCI Machine Learning Repository, five feature-selection algorithms were first applied, and a consensus retention rule reduced the input space to 23 robust variables. The balanced training data were then generated using SVM-SMOTE, SMOTE-Tomek, and SMOTE-ENN.

利用 UCI 机器学习库中的台湾破产预测数据集，研究首先应用了五种特征选择算法，并通过共识保留规则将输入空间缩减为 23 个稳健变量。随后，利用 SVM-SMOTE、SMOTE-Tomek 和 SMOTE-ENN 生成了平衡的训练数据。

Five ensemble machine learning classifiers, namely gradient boosting, extreme gradient boosting, histogram-based gradient boosting, LightGBM, and AdaBoost, were compared with five deep learning models, including RNN, LSTM, GRU, DNN, and MLP. In addition, hybrid stacking ensembles combined the five machine learning classifiers as base learners with each deep learning model as a meta-learner.

研究对比了五种集成机器学习分类器（梯度提升、极端梯度提升、基于直方图的梯度提升、LightGBM 和 AdaBoost）与五种深度学习模型（RNN、LSTM、GRU、DNN 和 MLP）。此外，混合堆叠集成模型将这五种机器学习分类器作为基学习器，并将每种深度学习模型作为元学习器进行了组合。

Model performance was assessed using accuracy, recall, specificity, G-mean, and ROC-AUC, while SHAP was used to explain feature contributions. The results show that resampling strategy materially shaped model behavior. SVM-SMOTE and SMOTE-Tomek favored accuracy and specificity, whereas SMOTE-ENN delivered stronger minority-class detection.

模型性能通过准确率、召回率、特异性、G-mean 和 ROC-AUC 进行评估，同时使用 SHAP 来解释特征贡献。结果表明，重采样策略对模型行为产生了实质性影响。SVM-SMOTE 和 SMOTE-Tomek 在准确率和特异性方面表现更优，而 SMOTE-ENN 在少数类别检测方面表现更强。

Among standalone models, the GRU with SMOTE-ENN achieved the best overall predictive balance, with recall of 0.8627, G-mean of 0.8517, and ROC-AUC of 0.9431. Among stacking ensembles, SMOTE-ENN with (GB+XGB+HGB+LGBM+AB)+LSTM provided the strongest compromise between sensitivity and specificity.

在独立模型中，采用 SMOTE-ENN 的 GRU 实现了最佳的整体预测平衡，其召回率为 0.8627，G-mean 为 0.8517，ROC-AUC 为 0.9431。在堆叠集成模型中，SMOTE-ENN 与 (GB+XGB+HGB+LGBM+AB)+LSTM 的组合在灵敏度和特异性之间提供了最强的平衡。

SHAP analysis identified leverage, profitability, solvency, and operational efficiency indicators as the most influential predictors of bankruptcy risk. These findings support more reliable and interpretable early warning systems for financially distressed firms.

SHAP 分析确定了杠杆率、盈利能力、偿债能力和运营效率指标是破产风险中最具影响力的预测因子。这些发现为财务困境企业的更可靠、更具可解释性的预警系统提供了支持。