Proposal of Automatic Credit Risk Evaluation

This book is focused on fuzzy logic application on the process automatic default client detection from the bank's credit risk management point of view. Based on contemporary Credit Risk Management information system analysis operating on manual default/non-default interpretation basis, the author suggests changing approach in a loan client evaluation and tried to automatize the process. Therefore static classification of different events must be put into wider context. The analysis proved, that the best solution is to change the whole model. Behavioral model has been employed in order to analyze discriminatory power of client related variables. Each significant variable has been taken into account in the transformation matrix setting up process. For further usage, each calculation is explained very clearly so it might be rewritten into the computer code right away. As a result, analysis shows that client behavioral is different even on such small market as is in the Czech Republic. Although the final results was not accurate enough to declare the methodology exact, the accuracy still might be considerably improved.