Bankruptcy Prediction Using Machine Learning: The Case of Slovakia

This research paper develops bankruptcy prediction models using machine learning techniques, specifically logistic regression and neural networks. Analyzing a dataset of 8159 companies from the Slovak Republic, the study highlights the superior performance of neural networks over logistic regression in terms of classification accuracy. Neural networks capture intricate patterns and relationships within the data, leveraging their flexibility and adaptability to achieve higher precision in predicting bankruptcies. Despite COVID-19 challenges, the models perform well due to early containment measures and support for small- and medium-sized companies. However, methodological limitations hinder individual bankruptcy identification, relying on financial metrics. The global impact of the COVID-19 pandemic, energy crisis, Ukrainian conflict, and high inflation persists. Future research should incorporate these factors into bankruptcy models, not only for the Slovak Republic but also for other transitioning economies. This exploration will enhance the understanding and accuracy of bankruptcy predictions.