Some Methods for Estimating Financial Risks in Banking

Abstract

This paper is devoted to the investigation of the possibilities of using neural networks, Altman model and linear regression for Z-score and Beta index prediction. These methods could be used for evaluating the probability of bankruptcy for companies-borrowers, individuals, portfolio of credits and the bank in whole. In this work Z-score was predicted for Bank of America based on Altman model and the best model for its forecasting was neural network backpropagation with the configuration (2,3,3,3,3,3,3,3,3,2). In future research is planning to make the further investigation of the possibilities to use the other data mining methods for evaluating the probability of risk and forecasting of expected losses after its appearance.