|
|
 |
| |
|
|
|
Development of GSTARIMA-ARCH Model for Rainfall Forecasting in Java Island using Big Data Analytics |
|
|
|
PP: 577-593 |
|
|
|
doi:10.18576/amis/190309
|
|
|
|
Author(s) |
|
|
|
Putri Monika,
Budi Nurani Ruchjana,
Atje Setiawan Abdullah,
Rahmat Budiarto,
|
|
|
|
Abstract |
|
|
| Spatio-Temporal (ST) model is developed with the combination of ARCH to overcome non-constant error variance through data analytics life cycle method. The subsequent model developed is Generalized Space-Time Autoregressive (GSTAR), which simultaneously considers spatial and temporal dependence in rainfall data. Following this process, GSTAR is combined with ARCH to overcome the assumption of heteroscedasticity in rainfall. Therefore, this research aimed to develop a combined GSTARIMA-ARCH to forecast rainfall on Java Island, which is characterized by high rainfall intensity. The methodology used in this research was analysis and modeling of GSTARIMA-ARCH in line with data analytics life cycle, particularly designed for handling Big Data in rainfall analysis. Consequently, the results showed that forecast produced by GSTARIMA-ARCH was more accurate than other conventional models. Moreover, this model could be used as a tool for decision-making by relevant agencies, as well as in the field of meteorology for exploring weather and rainfall. Finally, the model is also applicable in water resources management and natural disaster mitigation in Java Island. |
|
|
|
|
|
|
|
