<![CDATA[Bankruptcy prediction is a critical component of financial risk management, particularly in emerging capital markets such as Indonesia, where changes in corporate financial conditions can increase the need for reliable early-warning systems. This study develops an explainable ensemble machine learning framework that combines Random Forest and XGBoost to predict bankruptcy risk among companies listed on the Indonesia Stock Exchange (IDX) during 2019–2023. The analysis uses 200 companies, comprising 55 bankrupt and 145 non-bankrupt firms, represented by eight financial ratios. Five-fold stratified cross-validation is used to compare Logistic Regression, Decision Tree, Random Forest, XGBoost, and the proposed soft-voting Ensemble model using Accuracy, Precision, Recall, F1-Score, and AUC-ROC. The Ensemble model achieves the strongest overall performance, with an Accuracy of 91.2%, F1-Score of 89.8%, and AUC-ROC of 0.941. To improve interpretability, SHapley Additive exPlanations (SHAP) are applied to quantify the contribution of each financial ratio. Debt-to-Equity Ratio (DER) and Return on Assets (ROA) emerge as the most influential predictors, with mean absolute SHAP values of 0.312 and 0.287, respectively. The findings demonstrate that combining ensemble learning with SHAP can provide both strong predictive performance and transparent financial-risk insights, supporting data-driven early-warning applications for investors, auditors, and regulators in the Indonesian capital market.]]>
Copyrights © 2026
