Título: A Wavelet Network Classifier Applied to Financial Distress Prediction
Autores: Galvão, Roberto Kawakami H.; Becerra, Victor M.; Abou-Seada, Magda
Resumo: This work proposes a constructive method for building a wavelet network classifier. The network consists of a wavelet layer, which implements a nonlinear transformation in the input data, and a linear discriminant function, which carries out classif ication on the basis of the wavelet layer output. The proposed methodology is tested in a f inancial distress prediction problem involving British companies in the period 1997–2000. In this case study, the wavelet network was found to be a better classifier than a model obtained by linear discriminant analysis and a multi–layer perceptron trained with the Optimal Brain Damage technique.
Código DOI: 10.21528/CBRN2003-001
Artigo em PDF: 6CBRN_001.pdf
Arquivo BibTex: 6CBRN_001.bib