<![CDATA[The ability to accurately predict financial distress is crucial for State-Owned Enterprises (SOEs), given their strategic role in maintaining national economic stability. However, existing studies predominantly examine financial distress models in isolation and rely mainly on financial ratios, with limited attention to comparative evaluation under a unified machine learning framework and alternative input structures. This gap limits the understanding of how model performance may vary across different data representations. This study aims to evaluate and compare the predictive performance of four financial distress models Altman Z-Score, Springate S-Score, Zmijewski X-Score, and Grover G-Score by integrating them within a Naïve Bayes classification approach. Using a dataset of 20 Indonesian SOEs listed on the Indonesia Stock Exchange over the 2020–2023 period, this study applies a quantitative comparative method with two types of input variables, namely financial ratios and financial statement account balances. The results show that the Springate S-Score model demonstrates the highest predictive accuracy, achieving 95% when using financial ratios and 82.5% when using account balances. Overall, models based on financial ratios outperform those utilizing raw financial statement data, indicating that structured financial indicators provide more effective signals for classification. The main contribution of this study lies in providing a comprehensive and consistent comparison of multiple financial distress prediction models within a single probabilistic machine learning framework, while also highlighting the impact of different input variable structures on model performance. This study extends the financial distress literature by bridging traditional financial analysis and data mining approaches, and offers practical implications for developing more reliable early warning systems for financial distress in SOEs. Keywords : Financial Distress Prediction, Naïve Bayes, Machine Learning]]>
Copyrights © 2026
