The Optimal Determination Of Space Weight in Gstar Model by Using Cross-Correlation Inference
| Main Authors: | Suhartono, Suhartono, Subanar, Subanar |
|---|---|
| Format: | Article PeerReviewed application/pdf |
| Bahasa: | eng |
| Terbitan: |
, 2006
|
| Subjects: | |
| Online Access: |
https://repository.ugm.ac.id/32932/1/3.pdf https://repository.ugm.ac.id/32932/ |
Daftar Isi:
- Tlu:ai m of this paper is to discuss and develop the optima l determination of space we ight in GSTAR (Genera li zed Space-Time Au toreg ressive) mode l by applyi n g stat istica l inference of cross-correlation between l ocat ions (spaces) at the appropriate time lag. Our previous research sho''ed that the d irectly used of cross-correlat ion normalization as space weight give improper coefficient between locations in GSTAR model; i.e. these coefficients tend to be significant even though the true condition is insignificant. I n this p ;>c•i. we propose a statist ical test to va lidate the cross-correlat ion between locations that used as basic of space weight determination in GSTAR model. We focus on the GSTA R( 1 1) model and use three kinds relationship between locat ions as case studies. The results show that statistical inference process to validate cross-correlation between loca tions yields valid (unbiased) space weight estimates in GSTAR{I 1) model. In general, we can conclude that determination of space weight by using normalization of statistical infer ence to cross-correlation between locations at the appropriate time lag is the optimal procedure in GSTAR modeling.