Podejście wielomodelowe w prognozowaniu zagrożenia przedsiębiorstw upadłością w Polsce
The multiple-model approach in the prediction of company bankruptcy risk in Poland

Summary/Abstract: The subject matter of the paper fits into the mainstream of scientific research on an important economic problem, which is the bankruptcy of companies. Within the study of company bankruptcy risk, research is continued on the methodology for predicting the first and subsequent bankruptcies of companies. The aim of the paper is to present the results of empirical research on the usefulness of using the multiple-model approach in the prediction of company bankruptcy risk in Poland. The analysis was performed for balanced and unbalanced sets of enterprises from the industrial processing sector. The bagging, boosting, random subspaces and random forests methods were used in the research. The usefulness of the multiple-model approach for the prediction of bankruptcy was evaluated based on values of the classification efficiency measures for companies from the test set. The conclusions from the empirical research carried out indicate the usefulness of using the multiple-model approach in predicting the company bankruptcy risk.

• Details
• Contents