ORDER SELECTION OF AUTOREGRESSIVE MODEL
Order selection is an important step in autoregressive (AR) modeling. It may be executed through examining the stationary process and the behavior of residuals. In this thesis, the New Final Prediction Error (NFPE) approach was applied by calculating the expectation of prediction error. The best...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Subjects: | |
| Online Access: | https://digilib.itb.ac.id/gdl/view/33925 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Order selection is an important step in autoregressive (AR) modeling. It may be
executed through examining the stationary process and the behavior of residuals.
In this thesis, the New Final Prediction Error (NFPE) approach was applied by
calculating the expectation of prediction error. The best order was selected based on
the criteria of minimum prediction error values. As illustrations, several simulations
for dierent cases, which consider the combination of parameter values and the
order of AR model, were performed. For each case of these simulations, the Akaike
Information Criterion method (AIC), Bayesian Information Criterion (BIC), and
NFPE were compared for selecting the best order. It is obtained that NFPE give
the best predictive results than the other two methods. |
|---|