Research Article
Integrating Data into Machine Learning Models for Better Bankruptcy Prediction
@INPROCEEDINGS{10.4108/eai.28-4-2025.2358019, author={B. Rajani and R. Usha and Nambi Navya and Sadhama Thanuja and P Tarun and Sannithi Poojitha}, title={Integrating Data into Machine Learning Models for Better Bankruptcy Prediction}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part II}, publisher={EAI}, proceedings_a={ICITSM PART II}, year={2025}, month={10}, keywords={bankruptcy prediction hybrid machine learning imbalanced data set oversampling techniques ensemble deep learning financial risk management}, doi={10.4108/eai.28-4-2025.2358019} }- B. Rajani
R. Usha
Nambi Navya
Sadhama Thanuja
P Tarun
Sannithi Poojitha
Year: 2025
Integrating Data into Machine Learning Models for Better Bankruptcy Prediction
ICITSM PART II
EAI
DOI: 10.4108/eai.28-4-2025.2358019
Abstract
Bankruptcy prediction is important for financial health and risk control. We have improved the prediction accuracy using a hybrid of machine learning techniques AdaBoost and CatBoost on imbalanced Polish data set. The class imbalance and the noisy features along with the relatively insufficient bankrupt firm samples are challenges of our dataset. AdaBoost and CatBoost, powerful ensemble methods, work well with unbalanced data and interactions between features. It enhances prediction accuracy: reliable classification is provided even under adverse conditions. The experimental results demonstrate significant improvements in F1-score, recall, and precision compared to existing models. Importance analysis of distinguishing features reveals important financial indicators to the stakeholders. First, this flexible framework supports banks’ application of tools to address risks and improve economic resilience.
