Latuasan, Erza Claudia (2026) Peramalan Harga Beras Antar Wilayah Di Pulau Jawa Menggunakan Metode Generalized Space-Time Autoregressive With Exogenous Variable (GSTARX). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003221032-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003221032-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Beras merupakan komoditas pangan strategis di Indonesia karena menjadi makanan pokok bagi sebagian besar penduduk. Di Pulau Jawa, tiga provinsi utama, yakni Jawa Barat, Jawa Tengah, dan Jawa Timur, memegang peran penting sebagai episentrum produksi padi nasional. Harga beras di ketiga wilayah ini memiliki pola pergerakan yang dipengaruhi oleh faktor temporal dan spasial yang saling berkaitan. Ketidakpastian fluktuasi harga beras akibat dinamika pasar dan keterbatasan pasokan dapat memicu lonjakan inflasi, sehingga peramalan yang akurat menjadi kebutuhan strategis. Penelitian ini bertujuan untuk menganalisis karakteristik harga beras di tiga provinsi utama Pulau Jawa dan membangun model Generalized Space-Time Autoregressive with Exogenous Variable (GSTARX) yang mampu menangkap hubungan spasial dan temporal untuk memperoleh hasil peramalan di masa depan. Korelasi linier positif yang signifikan antarprovinsi mengindikasikan adanya efek dependensi antar lokasi. Pemodelan diawali dengan menggunakan regresi time series (TSR) untuk menangkap tren, pola musiman, dampak pandemi COVID-19, kebijakan Harga Eceran Tertinggi (HET), dan variasi kalender Hari Raya Idul Fitri. Sisaan dari model regresi tersebut kemudian dimodelkan ke dalam struktur spasio-temporal GSTAR. Model GSTARX terbaik diperoleh dari full model GSTARX(3) yang memiliki orde autoregresif (AR) p=3 dan orde spasial 1 dengan matriks bobot spasial seragam karena memenuhi asumsi white noise dan normalitas, serta menghasilkan kesalahan prediksi terkecil pada data out-sample. Peramalan menunjukkan adanya lonjakan harga yang signifikan hingga mencapai titik puncak tertinggi pada Maret 2026 akibat guncangan permintaan musiman momen Idul Fitri, dengan urutan tertinggi di Jawa Tengah, Jawa Barat, dan Jawa Timur, sebelum akhirnya melandai pada pertengahan tahun. Model ini dapat digunakan sebagai dasar strategis bagi Badan Pangan Nasional (Bapanas) dan Bulog dalam memantau stabilitas pangan regional.
=====================================================================================================================================
Rice is a strategic food commodity in Indonesia as it serves as the staple food for most of the population. On Java Island, three main provinces, namely West Java, Central Java, and East Java, play a vital role as the epicenter of national rice production. Rice prices in these three regions exhibit movement patterns influenced by interrelated temporal and spatial factors. Unpredictable fluctuations in rice prices driven by market dynamics and supply constraints can trigger inflation, making accurate forecasting a strategic necessity. This study aims to analyze the characteristics of rice prices in the three main provinces of Java Island and construct a Generalized Space-Time Autoregressive with Exogenous Variable (GSTARX) model capable of capturing spatial and temporal relationships to obtain future forecasts. A significant positive linear correlation between provinces indicates the presence of spatial dependency effects between locations. The modeling begins with Time Series Regression (TSR) to capture trends, seasonal patterns, the impact of the COVID-19 pandemic, Maximum Retail Price (HET) policies, and the calendar variation of Eid al-Fitr. The residuals from this regression model are then modeled into a GSTAR spatio-temporal structure. The best GSTARX model is obtained from the full model GSTARX(3) with a third-order autoregressive (AR(3)) and first-order spatial structure, using a uniform spatial weight matrix, as it satisfies the white noise and multivariate normality assumptions, while yielding the smallest prediction errors on out-sample data. Forecasting results indicate a significant price surge reaching its highest peak in March 2026 due to demand shocks during the Eid al-Fitr period, with the highest prices observed in Central Java, followed by West Java and East Java, before eventually leveling off by mid-year. This model can be utilized as a strategic foundation for the National Food Agency (Bapanas) and Bulog in monitoring regional food stability.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | GSTARX, Harga Beras, Variabel Eksogen, Peramalan, Spasial-Temporal, GSTARX, Rice Prices, Exogenous Variables, Forecasting, Spatio-Temporal |
| Subjects: | Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Erza Claudia Latuasan |
| Date Deposited: | 30 Jul 2026 06:47 |
| Last Modified: | 30 Jul 2026 06:47 |
| URI: | http://repository.its.ac.id/id/eprint/139948 |
Actions (login required)
 |
View Item |