Título: Misspecified Neural Network Models and Linear Time Series Forecasting
Autores: Pinto, Francisco Carlos de A.; Medeiros, Marcelo C.
Resumo: This paper studies the performance of neural networks estimated with Bayesian regularization to model and forecast time series where the data generating process is in fact a linear autoregression. A simulation experiment is carried out to compare the forecasts made by a correctly linear model and neural networks.
Palavras-chave:
Páginas: 10
Código DOI: 10.21528/lmln-vol3-no1-art5
Artigo em PDF: vol3-no1-art5.pdf
Arquivo BibTex: vol3-no1-art5.bib