A COMPARISON OF NEURAL NETWORKS AND ARCHGARCH MODELS TO PREDICT CHANGES IN SHARE PRICES. AN APPLICATION TO THE CASE OF STOCKS IN THE TELECOMUNICATIONS INDUSTRY
In recent years, there have been attempts to test the theory of market efficiency, using more efficient and accurate models to predict changes in the prices of various financial instruments. Actually there are two ways to predict such variations: parametric and nonparametric models. In the first gro...
| Autores principales: | , |
|---|---|
| Formato: | Artículo publishedVersion |
| Lenguaje: | Español |
| Publicado: |
Centro de Investigación en Métodos Cuantitativos Aplicados a la Economía y la Gestión (CMA)
2015
|
| Acceso en línea: | https://ojs.economicas.uba.ar/RIMF/article/view/1505 https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=modelfin&d=1505_oai |
| Aporte de: |
| Sumario: | In recent years, there have been attempts to test the theory of market efficiency, using more efficient and accurate models to predict changes in the prices of various financial instruments. Actually there are two ways to predict such variations: parametric and nonparametric models. In the first group there are various statistical-econometric models, while in the second there are artificial intelligence techniques as neural networks and genetic algorithms. The use of neural networks for predicting the behaviour of economic variables has increased greatly in recent years. This paper describes the design of solutions to forecast the share price of Telecom Argentina SA, which is listed on the Stock Exchange of Buenos Aires, in the period 2005-2012 from the use of a technique called Principal Component Analysis. The results are presented based on traditional models Arch-Garch and backpropagation networks. Additionally, a comparison between the methodologies is presented, considering the degree of prediction achieved. |
|---|