Pemodelan Peramalan Ekspor Kelapa Sawit dan Batu Bara sebagai Produk Komoditas Unggulan untuk Mendukung Perencanaan Ekonomi

Export and logistics planning in Bengkulu Province requires forecasting information that accounts for differences between the region’s leading commodities. This study aims to analyze historical patterns, select appropriate ARIMA models, and evaluate short-term forecasts of palm oil and coal export volumes to support operational planning at Pulau Baai Port. A quantitative approach was adopted, with each commodity modeled independently using the Autoregressive Integrated Moving Average (ARIMA) method. Monthly data from January 2019 to December 2024 were used for model development, while data from January to June 2025 were reserved for out-of-sample testing. Modeling followed the Box–Jenkins procedure, including stationarity testing, model identification, parameter estimation, diagnostic checking, and forecast accuracy evaluation using the Mean Absolute Percentage Error (MAPE). ARIMA(2,1,3) was selected for palm oil exports and ARIMA(0,1,1) for coal exports, yielding MAPE values of 1.79% and 2.13%, respectively. During the testing period, palm oil export forecasts showed an upward trend, whereas coal export forecasts remained relatively constant. Both models produced low percentage errors over the six-month testing period. These findings provide supporting information for storage capacity and shipping planning tailored to each commodity. Applications to subsequent periods require updated data and renewed evaluation, given the limited testing period and the univariate approach, which does not directly incorporate external factors.