Bankruptcy prediction, decision support for digital business

Limited liability companies can experience serious financial problems due to a lousy economy or unsuccessful business transactions. In the worst case, this can result in the company's bankruptcy. An indicator that will warn managers may be beneficial since the company may attempt to remedy its financial situation and take corrective actions. In reality, only a fraction of all companies face bankruptcy; therefore, we approach this issue as a classification of an imbalanced dataset. We work with the financial data of thousands of limited liability companies and try to predict bankruptcy several years before it happens.

We collected a dataset of limited liability companies that can be used to benchmark and validate methods for imbalanced learning and bankruptcy prediction. The dataset is publicly available on the Mendeley Data portal. A detailed dataset description can be found in our Small- and medium-enterprises bankruptcy dataset manuscript, published in the Data in Brief journal.