Application of Neuro-Fuzzy Approach in Predicting the Number of Bankruptcies of Legal Persons in the Czech Republic

Abstract:

This article deals with the application of the neuro-fuzzy approach in estimating the number of companies going bankrupt in the Czech Republic. The prediction is based on macroeconomic indicators from 2011 2016, namely inflation, interest rate and unemployment rate. Unlike statistical models, the neuro-fuzzy models have the advantage of rules made up directly from the used data, and therefore they enable modelling of complex, dynamic and non-linear problems. Based on the obtained results, it can be stated that the designed ANFIS (Adaptive Neuro- Fuzzy Inference System) is able to predict the number of financial failures of companies with sufficient accuracy and thus give a picture of the future market development.