Classification Models as Tools of Bankruptcy Prediction—Polish Experience

Abstract

The purpose of this paper is to present the results of extensive empirical research conducted in Poland on the capability of the bankruptcy prediction models most commonly used in practice to provide a correct classification. A main dataset and its variants, forming the basis for the conducted studies, were described. A comparative analysis of the most frequent bankruptcy prediction models was performed, based on the example of manufacturing companies in Poland. Empirical and theoretical distributions of financial ratios which are diagnostic variables in the presented models were examined. In summary, the conclusions were formulated allowing us to provide an answer to four basic questions raised in the introduction to the paper.