PENERAPAN MODEL AUTOREGRESSIVE FRACTIONALLY INTEGRATED MOVING AVERAGE DALAM MERAMALKAN NILAI TUKAR RUPIAH TERHADAP US DOLLAR (USD/IDR)

Ayu Fajar Rusadi (Program Studi Statistika, Universitas Lambung Mangkurat, Jl. A. Yani KM.35,5, Banjarbaru, Indonesia)
Yeni Rahkmawati (Program Studi Statistika, Universitas Lambung Mangkurat, Jl. A. Yani KM.35,5, Banjarbaru, Indonesia)
Fitri Handayani (Badan Pusat Statistik Provinsi Kalimantan Selatan, Jl. Soekarno Hatta/Trikora No 7 Banjarbaru, Indonesia)

Jurnal Gaussian
Vol 14, No 2 (2025): Jurnal Gaussian

The Autoregressive Fractionally Integrated Moving Average (ARFIMA) model is employed to analyze data exhibiting long memory characteristics using a fractional differencing coefficient. This study differs from previous research, as no existing studies have been found that discuss the forecasting of the rupiah exchange rate against the United States Dollar (USD) using recent data and the ARFIMA model. This study examines the daily exchange rate of the Rupiah against the United States Dollar (USD/IDR) from January 2023 to March 2025, totalling 535 observations. The results indicate a general weakening trend of the Rupiah during this period. The Hurst exponent (H = 0.8372489), which falls within the range 0.5 < H < 1, confirms the presence of long memory properties. The best fitting ARFIMA model is identified as ARFIMA(2,0.3372489,1), with a BIC value of 4387.69. The model equation is: . Forecasting results show a downward trend, suggesting potential appreciation of the rupiah against the USD in the future. These findings provide valuable insights into exchange rate dynamics and have important implications for economic planning and policy in Indonesia.