Choosing a method for predicting economic performance of companies

Abstract

This paper reports on the efforts to find a method for predicting economic results of companies. The input data files consist of 93 profitable companies and 93 bankrupt firms. From the total number of 93 firms in both categories, data of 72 firms served for establishing a classification criterion, and for the rest of 21 firms, a prognosis of their economic development was performed. The classification criterion for prognosticating the future economic development has been established by applications of discriminate analysis, logit analysis, and artificial neural network methods. The application of artificial neural networks has provided for better classification accuracies of 90,48 % for successful firms, and 100 % for bankrupt firms.

Keywords

• prediction
• discriminant analysis
• logit analysis
• artificial neural networks
• classification accuracies