Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Similars in SciELO
Share
Tékhne - Revista de Estudos Politécnicos
Print version ISSN 1645-9911
Abstract
PEREIRA, José Manuel; DOMINGUEZ, Miguel Á. Crespo and OCEJO, José L. Sáez. Modelos de previsão do fracasso empresarial: aspectos a considerar . Tékhne [online]. 2007, n.7, pp.111-148. ISSN 1645-9911.
Prediction of corporate bankruptcy is a phenomenon of increasing interest to investors, creditors, borrowing firms, and governments alike. Since the seminal work of Beaver (1966) different techniques have been used: discriminant analysis, logit, probit, neural networks, decision trees, genetic algorithms, rough sets, and some other techniques. Our intent is to provide a comparison of the most popular methods, analysing their own strengths and weaknesses and their applicability to potential users. We find that predictive accuracies of different models seem to be generally comparable and the use of discriminant analysis and logit models dominates the research. In general the neural networks models are the most difficult for the users.
Keywords : Bankruptcy; Discriminant Analysis; Logit; Probit; Neural Networks; Decision Trees.