Modelování a predikce pravděpodobnosti selhání pro retailové úvěry

The aim of this thesis is to model and estimate the probability of default for retail loans using several selected methods in accordance with the methodology IFRS 9. These methods are logistic regression, decision trees, survival analysis and absorbing Markov chains. The theoretical chapters involve topics related to the probability of default in the context of credit risk management, associated financial theory, empirical research and regulation of the banking environment. The main emphasis is there put on the description of individual modelling methods. The empirical chapters involve complete data modelling processes according to data mining methodology including the estimation of the probability of default for all methods.