Serviços Personalizados
Journal
Artigo
Indicadores
- Citado por SciELO
- Acessos
Links relacionados
- Similares em SciELO
Compartilhar
Ingeniería, investigación y tecnología
versão On-line ISSN 2594-0732versão impressa ISSN 1405-7743
Resumo
ACOSTA-CERVANTES, M.C.; VILLARREAL-MARROQUIN, M.G. e CABRERA-RIOS, M.. Validation Study on Artificial Neural Network-Based Selection of Time Series Forecasting Techniques. Ing. invest. y tecnol. [online]. 2013, vol.14, n.1, pp.53-63. ISSN 2594-0732.
In this paper, a validation study for a method geared towards the selection of forecasting techniques for time series is presented. The proposed method makes use of artificial neural networks to predict the performance of several statistics-based forecasting techniques to help select the potential best one. Eighteen time series with real data related to economic activities in the state of Tamaulipas were used for validation purposes. The results indicate that the proposed method is sufficiently reliable to become a useful resource for people with modest level of training in statistics. It is also proposed that the method be tabulated for convenient access.
Palavras-chave : Artificial Neural Network (ANN); forecasting methods; time series.