Comparing the bank failure prediction performance of neural networks and support vector machines: the Turkish case

摘要: AbstractExperience from the banking crises during past two decades suggest that advanced prediction models are needed for helping prevent bank failures. This paper compares ability of artificial neural networks and support vector machines in predicting Although have widely been applied complex problems business, literature utilizing is relatively narrow their capability failures not very familiar. In this paper, these intelligent techniques to a dataset Turkish commercial banks. Empirical findings show although performance can be considered as satisfactory, slightly better predictive than machines. addition, different types error each model also indicate network predictors.