Asrirawan, Asrirawan (2014) Model Generalized Seasonal Autoregressive Integrated Moving Average (GSARIMA) Untuk Peramalan Jumlah Penderita Demam Berdarah Dengue (DBD) Di Kota Surabaya. Masters thesis, Insititut Teknologi Sepuluh Nopember.
![[thumbnail of 1312201028_Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
1312201028_Master_Thesis.pdf Download (2MB) |
Abstract
Salah satu topik utama dalam kajian pemodelan peramalan deret waktu (time series) pada tiga dekade terakhir ini adalah peramalan data jumlahan (count data). Peramalan data jumlahan berbasis model stokastik masih belum banyak dilakukan dan selama ini masih menggunakan distribusi Gaussian (normal). Salah satu model stokastik dan non-Gaussian untuk peramalan data jumlahan adalah model generalized autoregressive moving average (GARMA). Model GARMA menghubungkan komponen ARMA dengan variabel prediktor ke transformasi parameter rata-rata dari distribusi data dengan menggunakan fungsi link (link function) tetapi tidak melibatkan efek nonstasioner dan musiman. Model generalized seasonal autoregressive integrated moving average (GSARIMA) merupakan model pengembangan dari GARMA dengan melibatkan efek nonstasioner dan musiman. Estimasi model GSARIMA menggunakan pendekatan iteratively reweighted least square (IRLS). Model GSARIMA diterapkan pada data simulasi dan kasus demam berdarah dengue (DBD) di kota Surabaya. Selain itu, model GSARIMA dibandingkan dengan model seasonal autoregressive integrated moving average (SARIMA). Hasil analisis pada data simulasi dan studi kasus menunjukkan bahwa model GSARIMA lebih baik dibanding model SARIMA dengan menggunakan nilai AIC dan mean absolute relative error MARE.
=============================================================================================================================
One of the main topics in the study of time series forecasting models in the last three decades is count data forecasting. The forecasting of count data based on stochastic model has not been applied and its normally used Gaussian distribution. One of stochastic models for count data forecasting, which has non-Gaussian (negative binomial) distribution, is generalized autoregressive moving average (GARMA) models. GARMA models relate ARMA components and predictor variable via a transformation of mean parameters of the data distribution using a link function but these models do not consider nonstationary and seasonal effects. Generalized seasonal autoregressive integrated moving average (GSARIMA) models are an extended version of GARMA involving nonstationary and seasonal effects. The estimation of GSARIMA models could be done by an approach iteratively reweighted least square (IRLS). This paper simulates GSARIMA models and applies it to dengue hemorrhagic fever (DHF) in Surabaya. Moreover, the forecast accuracy of GSARIMA Models is compared with seasonal autoregressive integrated moving average (SARIMA) models. The results show that GSARIMA models is better than the SARIMA by using AIC and mean absolute relative error (MARE) value both simulation and DHF data.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTSt 519.54 Asr m-1, 2014 |
| Uncontrolled Keywords: | Demam Berdarah Dengue, GSARIMA, IRLS, binomial negatif, SARIMA, Dengue Hemorrhagic Fever |
| Subjects: | Q Science > QA Mathematics > QA280 Box-Jenkins forecasting |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Mr. Marsudiyana - |
| Date Deposited: | 08 Jan 2024 06:18 |
| Last Modified: | 08 Jan 2024 06:18 |
| URI: | http://repository.its.ac.id/id/eprint/105399 |
Actions (login required)
 |
View Item |