Publicação
The efficiency of bankruptcy predictive models - genetic algorithms approach
| Resumo: | The present dissertation evaluates the contribution of genetic algorithms to improve the performance of bankruptcy prediction models. The state-of-the-art points to a better performance of MDA (Multiple Discriminant Analysis)-based models, which, since 1968, are the most applied in the field of bankruptcy prediction. These models usually recur to ratios commonly used in financial analysis. From the comparative study of (1) logistic regression-based models with the forward stepwise method for feature selection, (2) Altman's Z-Score model (Edward I. Altman, 1983) based on MDA and (3) logistic regression with the contribution of genetic algorithms for variable selection, a clear predominance of the efficiency revealed by the former models can be observed. These new models were developed using 1887 ratios generated a posteriori from 66 known variables, derived from the accounting, financial, operating, and macroeconomic analysis of firms. New models are thus presented, which are very promising for predicting bankruptcy in the medium to long term, in the context of increasing instability surrounding firms for different countries and sectors. |
|---|---|
| Autores principais: | Antão, Mário Manuel Neto |
| Assunto: | Genetic algorithm Logit Bankruptcy MDA Z-Score Logistic Regression Algoritmos Genéticos Falência Regressão Logística |
| Ano: | 2020 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | The present dissertation evaluates the contribution of genetic algorithms to improve the performance of bankruptcy prediction models. The state-of-the-art points to a better performance of MDA (Multiple Discriminant Analysis)-based models, which, since 1968, are the most applied in the field of bankruptcy prediction. These models usually recur to ratios commonly used in financial analysis. From the comparative study of (1) logistic regression-based models with the forward stepwise method for feature selection, (2) Altman's Z-Score model (Edward I. Altman, 1983) based on MDA and (3) logistic regression with the contribution of genetic algorithms for variable selection, a clear predominance of the efficiency revealed by the former models can be observed. These new models were developed using 1887 ratios generated a posteriori from 66 known variables, derived from the accounting, financial, operating, and macroeconomic analysis of firms. New models are thus presented, which are very promising for predicting bankruptcy in the medium to long term, in the context of increasing instability surrounding firms for different countries and sectors. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.