PERAMALAN NILAI TUKAR PETANI PROVINSI ACEH: DITINJAU DENGAN METODE DOUBLE EXPONENTIAL SMOOTHING DAN HOLT WINTER

Abstract

This research aimed to compare the performance of the Double Exponential Smoothing and Holt-Winter methods in forecasting the Farmer Exchange Rate (NTP) for the Horticulture Farmer Subsector in Aceh Province. Secondary data on the NTP for the Aceh Horticulture Subsector from 2020-2023 was obtained from the Aceh Provincial Central Statistics Agency (BPS) website. The data processing was carried out using the R software. The results showed that the Double Exponential Smoothing method had a Root Mean Squared Error (RMSE) of 3.413201, a Mean Absolute Percentage Error (MAPE) of 2.85835, and a Mean Percentage Error (MPE) of -0.5980113. Meanwhile, the Holt-Winter's Additive method had an RMSE of 4.055652, a MAPE of 3.055992, and an MPE of 0.2534276. The best forecasting method was selected based on the smallest RMSE, MAPE, and MPE values, which was Double Exponential Smoothing. The 12-month forecast using the Double Exponential Smoothing method indicated that the lowest NTP would occur in January 2024 at 106.7195, while the highest NTP would be in December 2024 at 113.6443

Keywords

• double exponential smoothing, holt winter and farmer exchange rate
• nilai tukar petani