IMPLEMENTASI METODE BAYESIAN STRUCTURAL TIME SERIES (BSTS) DALAM MEMPREDIKSI NILAI TUKAR YUAN TERHADAP RUPIAH

FARAH FAUZIAH NOVIANTI, . (2026) IMPLEMENTASI METODE BAYESIAN STRUCTURAL TIME SERIES (BSTS) DALAM MEMPREDIKSI NILAI TUKAR YUAN TERHADAP RUPIAH. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (2MB)
	Text BAB 1.pdf Download (510kB)
	Text BAB 2.pdf Restricted to Registered users only Download (784kB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (826kB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (647kB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (500kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (510kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (602kB) | Request a copy

Abstract

Penelitian ini menganalisis peramalan nilai tukar yuan terhadap rupiah menggunakan metode Bayesian Structural Time Series (BSTS). Nilai tukar mata uang adalah indikator penting bagi stabilitas perdagangan internasional, arus modal, dan formulasi kebijakan moneter. Model peramalan deret waktu konvensional seperti ARIMA dibatasi oleh kebutuhan akan data yang stasioner dan ketidakmampuannya mengakomodasi variabel eksternal secara fleksibel. Penelitian ini mengimplementasikan BSTS, yang mengintegrasikan komponen struktural dari deret waktu dengan kerangka Bayesian, sehingga memungkinkan estimasi fleksibel tanpa persyaratan stasioneritas dan akomodasi variabel eksternal melalui spike-and-slab prior. Data empiris mencakup observasi bulanan periode Januari 2009–Desember 2024 (n = 192) dengan lima faktor eksternal: suku bunga, inflasi, neraca perdagangan, harga minyak mentah dunia, dan harga emas dunia. Penelitian ini membentuk 18 spesifikasi model BSTS melalui kombinasi enam struktur model yang berbeda, yaitu variasi komponen tren dan musiman, dengan tiga jumlah iterasi MCMC (1000, 2000, dan 5000). Seluruh model kemudian dievaluasi berdasarkan kemampuan fit dalam-sampel. Model optimal (level lokal tanpa komponen musiman) mencapai R-Squared dalam-sampel sebesar 0,977346 dengan 2000 iterasi MCMC. Berdasarkan posterior inclusion probability, suku bunga menunjukkan pengaruh paling signifikan (PIP = 0,680), diikuti harga emas dunia (PIP = 0,498) dan inflasi (PIP = 0,399). Pada data uji (n = 39), model menghasilkan MAPE 2,82%, MAE 60,70 Rp/Yuan, dan RMSE 77,60 Rp/Yuan. Hasil ini menunjukkan bahwa BSTS merupakan metode peramalan yang valid untuk nilai tukar mata uang dengan mengakomodasi pengaruh faktor-faktor makroekonomi.***** This research analyzes the forecasting of the yuan-to-rupiah exchange rate using the Bayesian Structural Time Series (BSTS) method. Exchange rates serve as important indicators for international trade stability, capital flows, and monetary policy formulation. Conventional time series forecasting models such as ARIMA are limited by their requirement for stationary data and their inability to flexibly accommodate external variables. This research implements BSTS, which integrates structural time series components with a Bayesian framework, thereby enabling flexible estimation without stationarity requirements and accommodation of external variables through spike-and-slab priors. The empirical data comprises monthly observations from January 2009 to December 2024 (n = 192) with five external factors: interest rates, inflation, trade balance, global crude oil prices, and global gold prices. This research constructs 18 BSTS model specifications through combinations of six different structures—variations in trend and seasonal components—with three MCMC iteration counts (1000, 2000, and 5000). These models are then evaluated based on their in-sample fit performance. The optimal model (local level without seasonal component) achieves an in-sample R-Squared of 0.977346 with 2000 MCMC iterations. Based on posterior inclusion probability, interest rates demonstrate the most significant influence (PIP = 0.680), followed by global gold prices (PIP = 0.498) and inflation (PIP = 0.399). On the test data (n = 39), the model yields a MAPE of 2.82%, MAE of 60.70 Rp/Yuan, and RMSE of 77.60 Rp/Yuan. These results indicate that BSTS is a valid forecasting method for exchange rates while accommodating the influence of macroeconomic factors.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Drs. Sudarwanto, M.Si., DEA ; 2). Ibnu Hadi, M.Si.
Subjects:	Sains > Matematika
Divisions:	FMIPA > S1 Matematika
Depositing User:	Farah Fauziah Novianti .
Date Deposited:	04 Mar 2026 02:46
Last Modified:	04 Mar 2026 02:46
URI:	http://repository.unj.ac.id/id/eprint/65943

Actions (login required)

View Item