CEU Electronic Theses and Dissertations, 2025
| Author | Lleshi, Armando |
|---|---|
| Title | Predicting Financial Distress through the Development of a Hybrid Data-Driven Model |
| Summary | This project develops a predictive model to estimate the probability of financial distress among Hungarian firms using a hybrid approach that combines the Altman Z-score with logistic regression. Traditional risk assessment tools often miss early warning signs or lack scalability, especially in emerging markets. To address this issue, this model was created and trained on a dataset of over 23,000 private Hungarian companies, primarily in the agricultural sector. Distress was proxied using indicators like negative equity, persistent losses, and inactivity. Five key financial ratios were selected, tested for multicollinearity, and used as inputs in a logistic regression framework. The resulting model demonstrated high discriminative power and strong calibration. While an alternative LASSO model showed better generalization on new data, logistic regression offered greater interpretability. To enhance usability, the model was deployed through a web-based interface, enabling practical application in real-world financial risk monitoring. |
| Supervisor | Schindele, Ibolya |
| Department | Economics MSc |
| Full text | https://www.etd.ceu.edu/2025/lleshi_armando.pdf |
Visit the CEU Library.
© 2007-2025, Central European University