Research

Bankruptcy Prediction and Decision Support Systems for Digital Businesses

Bankruptcy

Big Data

Data Mining

Decision Support Systems

Imbalanced Data

Limited liability companies can face serious financial difficulties due to adverse economic conditions or unsuccessful business activities. In severe cases, these challenges may lead to bankruptcy. Early identification of financial distress is therefore essential, as it enables managers and stakeholders to take timely corrective actions and reduce potential losses.

Our research focuses on developing predictive models that can detect signs of bankruptcy several years in advance. Since only a small proportion of companies actually go bankrupt, the problem involves highly imbalanced data — a common challenge in real-world financial analytics. We address this by applying advanced classification and imbalanced learning techniques to financial data from thousands of companies.

To support research transparency and benchmarking, we compiled a comprehensive dataset of small and medium-sized enterprises (SMEs). The dataset is publicly available on the Mendeley Data repository and is documented in detail in our Small- and Medium-Enterprises Bankruptcy Dataset publication in the journal Data in Brief.

Related publications:

2025

P. Gnip, P. Drotár, R. Kanász, M. Zoričák

A Deep Ensemble Learning Approach for Imbalanced Data in Bankruptcy Prediction

Journal: 2025 IEEE Symposium on Computational Intelligence for Financial Engineering and Economics Cifer 2025.

P. Gnip, R. Kanász, M. Zoričák, P. Drotár

An experimental survey of imbalanced learning algorithms for bankruptcy prediction

Journal: Artificial Intelligence Review

2024

R. Kanász, P. Drotár, P. Gnip, M. Zoričák

Clash of titans on imbalanced data: TabNet vs XGBoost

Journal: 2024 IEEE Conference on Artificial Intelligence.

2023

R. Kanász, P. Gnip, M. Zoričák, P. Drotár

Bankruptcy prediction using ensemble of autoencoders optimized by genetic algorithm

Journal: PeerJ. Computer science.

2021

P. Gnip, L. Vokorokos, P. Drotár

Selective oversampling approach for strongly imbalanced data

Journal: PeerJ. Computer science.

2020

M. Zoričák, P. Gnip, P. Drotár, V. Gazda

Bankruptcy prediction for small- and medium-sized companies using severely imbalanced datasets

Journal: Economic Modelling

2019

P. Drotár, P. Gnip, M. Zoričák, V. Gazda

Small- and medium-enterprises bankruptcy dataset

Journal: Data in Brief.

2018

P. Drotár, P. Gnip, M. Zoričák, V. Gazda

Single-Class Bankruptcy Prediction Based on the Data from Annual Reports

Journal: Intelligent data engineering and automated learning