Aplicação de Redes Neurais Polinomiais na Previsão do Ibovespa e Merval
This article analyses the efficiency of Group Method of Data Handling (GMDH) polynomial neural networks when anticipating return, on a monthly basis, on the return of the main Brazilian (Ibovespa) and Argentinean (Merval) market indicators. Initially, in order to determine the exogenous variable, we...
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | PT |
| Publicado: |
Universidade Regional do Noroeste do Estado do Rio Grande do Sul
2011
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/e60de56adc4248a1b3a2c91ad88aa845 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | This article analyses the efficiency of Group Method of Data Handling (GMDH) polynomial neural networks when anticipating return, on a monthly basis, on the return of the main Brazilian (Ibovespa) and Argentinean (Merval) market indicators. Initially, in order to determine the exogenous variable, we calculated the logarithmical return on each index. Afterwards, in order to determine the endogenous variables, we have performed t-1, t-2 and t-3 lags on the exogenous variable. We computed up to nine front fed layers. Results suggest some predictability on both markets, denoting some inefficiency. Inefficiency, especially on the Argentinean market, is validated by the additional causality Granger tests that demonstrate the influence of the São Paulo Stock Market over the Buenos Aires Stock Market and no such influence the other way round. |
|---|