PREDICTION OF HURRICANE KATRINA LANDFALL WITH SPACE TIME GSTAR(p; )-ARCH(1) MODEL
Space-time analysis is used to modeling data with time and spatial dependency. One of space – time analysis model is Generalized Space Time Autoregressive (GSTAR) with assumption of constant residual variance. In this final project, GSTAR model will be constructed with residual that has inconstant r...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Subjects: | |
| Online Access: | https://digilib.itb.ac.id/gdl/view/33847 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Space-time analysis is used to modeling data with time and spatial dependency. One of space – time analysis model is Generalized Space Time Autoregressive (GSTAR) with assumption of constant residual variance. In this final project, GSTAR model will be constructed with residual that has inconstant residual variance or has a heteroscedastic effect. Autoregressive Conditional Heteroscedastic (ARCH) is use to model the unconstant variance of the residual. GSTAR –ARCH model parameter will be estimated using Generalized Least Square (GLS) method to get efficient parameter. GSTAR –ARCH model will be applied in daily average wind speed data for New Orleans, Florida and Mississippi to predict Hurricane Katrina landfall that had happened on 2005. The overall modeling shows that by using GSTAR(3;0,0,1)-ARCH(1) model, Hurricane Katrina is predicted to has a landfall on September 1st which three days later than the actual date. |
|---|