Advances and Applications in Statistics
Volume 48, Issue 3, Pages 169 - 183
(March 2016) http://dx.doi.org/10.17654/AS048030169 |
|
THE SHQC CRITERION FOR ORDER SELECTION IN AUTOREGRESSIVE MODELS
Enas Gawdat Yehia
|
Abstract: We propose the SHQC as a new order selection criterion based on both two information criteria, Schwarz information criterion (SIC) and Hannan-Quinn information criterion (HQC) for the selection in autoregressive models. The performance of our proposed criterion was investigated through a simulation study by comparing it with other well known order selection criteria. The simulation results show that the proposed criterion, SHQC is much better than other existing criteria in identifying the correct model order for moderate to large samples. Moreover, as the sample size increases, the SHQC converges to the true order which confirms the consistency of it and the probabilities of both under and over fitting of it are always the least among other criteria for all cases. Therefore, we can use the SHQC criterion as a safe alternative to any criterion. |
Keywords and phrases: autoregressive models, order selection, information criteria, SIC, HQC, simulation. |
|
Number of Downloads: 380 | Number of Views: 1300 |
|