Wybór rzędu autoregresji w zależności od parametrów modelu generującego
Autoregressive order selection depending on parameters of generating model

Summary/Abstract: The paper shows the performance of different criteria selecting the autoregression order: AIC, BIC, HQ and sequential rule of testing the significance of the last coefficient) assuming different values of autoregression coeficients and different sample sizes. It has been indicated that corrected AIC (AICC), rarely used in Polish studies, is very useful in selecting the autoregression order particularly in small samples. It has been emphasized that the issue of order selection may be considered in two contexts, namely in the context of selecting the true order of autoregressive model (true data generating model) and in choosing the best forecasting model. These considerations have been illustrated by the results of simulation study.

• Details
• Contents