Peramalan Harga Minyak Mentah Dunia Menggunakan Metode Radial Basis Function Neural Network
DOI:
https://doi.org/10.24036/r4mr9m27Keywords:
Crude Oil Price, Forecasting, RBFNN, MAPE, Artifical Neural NetworkAbstract
Crude oil prices are a key indicator of global economic stability because they affect the energy, transportation, and fiscal policy sectors. This study applies the Radial Basis Function Neural Network
(RBFNN) method to forecast monthly crude oil prices from February 2000 to July 2025 using normalized data. By varying the center percentage, the best model configuration was obtained at lags 1, 2, 53,
and 54, using 80% training data and 14% center. This combination yielded a MAPE value of 6.48% (very good category). The model was then used to predict crude oil prices up to July 2026, showing a
relatively stable price trend between USD 53.39 and USD 66.28 per barrel. These results indicate that RBFNN is a highly accurate method for supporting energy policy formulation and economic planning
