PERBANDINGAN RECURRENT FORECASTING DAN VECTOR FORECASTING PADA METODE HYBRID SSA-ARIMA DALAM PERAMALAN NILAI TUKAR DOLAR AMERIKA TERHADAP RUPIAH

AVISA SALSABILA PUTRI HARSANTI, . (2026) PERBANDINGAN RECURRENT FORECASTING DAN VECTOR FORECASTING PADA METODE HYBRID SSA-ARIMA DALAM PERAMALAN NILAI TUKAR DOLAR AMERIKA TERHADAP RUPIAH. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (1MB)
	Text BAB 1.pdf Download (3MB)
	Text BAB 2.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text Daftar Pustaka.pdf Download (3MB)
	Text Lampiran.pdf Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Nilai tukar Dolar Amerika Serikat terhadap Rupiah (USD/IDR) memiliki tingkat volatilitas yang tinggi akibat kerentanannya terhadap guncangan makroekonomi dan eskalasi geopolitik global. Oleh karena itu, penelitian ini bertujuan untuk memodelkan dan memproyeksikan nilai tukar tersebut menggunakan metode Hybrid Singular Spectrum Analysis (SSA) - Autoregressive Integrated Moving Average (ARIMA), dengan membandingkan keandalan algoritma Recurrent Forecasting dan Vector Forecasting pada tahap ekstrapolasi. Data yang digunakan adalah harga penutupan harian (daily close) USD/IDR selama periode 1 Januari 2025 hingga 31 Mei 2026. Hasil komputasi pada tahap pengujian (testing) membuktikan bahwa model Hybrid SSA (Recurrent) - ARIMA memiliki tingkat akurasi yang lebih optimal dengan Mean Absolute Percentage Error (MAPE) sebesar 1,1349% dan Root Mean Square Error (RMSE) sebesar 284,8904. Kinerja ini mengungguli varian Vector yang menghasilkan tingkat galat sebesar 1,3298% (MAPE) dan 333,0925 (RMSE). Selain terbukti sangat presisi dalam memodelkan pola data historis pada kondisi volatilitas normal, hasil proyeksi ekstrapolasi untuk 30 hari ke depan membuktikan bahwa algoritma Recurrent memiliki stabilitas komputasi yang sangat tangguh. Algoritma ini secara efektif mampu mengekstrapolasi momentum tren eskalasi di akhir periode observasi secara proporsional. Temuan ini menyimpulkan bahwa model Hybrid SSA (Recurrent) - ARIMA merupakan instrumen prediktif yang andal dalam menghasilkan lintasan taksiran masa depan yang logis, realistis, dan terukur, tanpa memunculkan bias lonjakan eksponensial yang tak terkendali. **** The exchange rate of the United States Dollar to the Indonesian Rupiah (USD/IDR) possesses a high level of volatility due to its vulnerability to macroeconomic shocks and global geopolitical escalations. Therefore, this study aims to model and project the exchange rate using the Hybrid Singular Spectrum Analysis (SSA) - Autoregressive Integrated Moving Average (ARIMA) method, by comparing the reliability of the Recurrent Forecasting and Vector Forecasting algorithms at the extrapolation stage. Utilizing daily close price data of USD/IDR over the period of January 1, 2025, to May 31, 2026, the computational results in the testing phase prove that the Hybrid SSA (Recurrent) - ARIMA model has a more optimal accuracy level with a Mean Absolute Percentage Error (MAPE) of 1.1349% and a Root Mean Square Error (RMSE) of 284.8904. This performance outperforms the Vector variant, which yields an error rate of 1.3298% (MAPE) and 333.0925 (RMSE). In addition to being highly precise in modeling historical data patterns under normal volatility conditions, the extrapolation projection results for the next 30 days prove that the Recurrent algorithm possesses highly robust computational stability. This algorithm is effectively capable of extrapolating the escalating trend momentum at the end of the observation period proportionally. These findings conclude that the Hybrid SSA (Recurrent) - ARIMA model is a reliable predictive instrument in generating logical, realistic, and measurable future estimation trajectories, without giving rise to an uncontrolled exponential surge bias.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Ibnu Hadi, M.Si. ; 2). Devi Eka Wardani Meganingtyas, S.Pd., M.Si.
Subjects:	Sains > Matematika
Divisions:	FMIPA > S1 Matematika
Depositing User:	Avisa Salsabila Putri Harsanti .
Date Deposited:	21 Aug 2026 01:25
Last Modified:	21 Aug 2026 01:25
URI:	http://repository.unj.ac.id/id/eprint/72692

Actions (login required)

View Item