Affiliation:
1. Kaunas University of Technology
Abstract
This research investigates the possibility to classify the companies into default and non-default groups analyzing the financial data of 1 year. The developed statistical model enables banks to predict the default of new companies that have no sufficient financial information for the credit risk assessment using other models. The classification and regression tree predicts the default of companies with the 96 % probability. The complementary analysis the financial data of 2 years by probit model allows to increase the classification accuracy to 99 %.
Key words: bank, classification, credit risk, statistical analysis.
Publisher
Vytautas Magnus University