HOW DO DIFFERENT TIME SPANS AFFECT THE PREDICTION ACCURACY OF BUSINESS FAILURE

Abstract

The prediction of business failure has been widely studied by many authors. Most of the studies focused on improve the results by applying new methodologies or by using more suitable financial information. This study aims to analyze the impact of the input data timeframe on the prediction accuracy of business failure. Using an artificial neural network, the self-organizing maps (SOM), we compare the results obtained by using 9, 6 and 3 years of input data. We concluded that the 3-year case provides a better global results despite of the 6-year case presents the lowest error type I.