<![CDATA[This study aims to systematize the development of machine learning applications for predicting corporate bankruptcy risk using a PRISMA-based Systematic Literature Review (SLR). A total of 17 articles published between 2015 and 2025 were analyzed to map research trends, compare algorithm performance, and evaluate the role of financial and non-financial data in prediction models. The findings indicate a clear shift from traditional statistical approaches toward machine learning algorithms such as SVM, Random Forest, ANN, and Deep Neural Networks, which consistently demonstrate higher accuracy across various countries and industries. The integration of Natural Language Processing (NLP), particularly annual report text analysis using BERT, enhances early detection of financial distress. However, challenges remain, including imbalanced data, overfitting risks, and limited model interpretability. These insights contribute to the development of more adaptive bankruptcy prediction models and highlight the importance of incorporating Explainable AI (XAI) to improve model transparency and reliability.]]>
Copyrights © 2026
