Zhodnocení spolehlivosti predikčních modelů na příkladu obchodních společností v České republice

This dissertation analyses and evaluates the predictive power of selected bankruptcy (predictive) models designed to forecast financial distress or impending bankruptcy of companies. The aim of this thesis is to verify whether the selected predictive models can reliably and in a timely manner predict the bankruptcy of companies that actually went bankrupt between 2021 and 2025. Another objective is to identify indicators that may have contributed to a potential misclassification to a greater extent, thereby constituting a Type I error. The theoretical section focuses on defining the concept of financial distress and insolvency and on the characteristics of the selected predictive models, which will subsequently be tested. In the practical section, these models are applied on companies trading goods in Czechia that have gone into insolvency over the past five years. The results of the analysis show that Taffler’s model and the IN05 Index achieved the highest classification accuracies; overall, the use of predictive models can be recommended.