Indonesian Crude Oil Price (ICP) Prediction Using Support Vector Regression Algorithm

Indonesian Crude Oil Price (ICP) Prediction Using Support Vector Regression Algorithm

ICP;prediction;SVR;RMSE;,MAPE

Indonesian crude oil prices (ICP) experience fluctuating movements, influenced by several factors and other conditions that make ICP prices difficult to predict. ICP priceprediction can be done with the Support Vector Regression (SVR) method. The information utilized originates from the Ministry of Energy and Mineral Resources' official website, specifically focusing oncrude oil pricing data for six primary types of crudeoil: SLC, Attaka, Duri, Belida, Banyuand SC. The data applied covers the time framefrom January 2018 to August 2023. The forecast of the ICP relies on the dateBrent variable and the Alpha factor through the useof support vector regression(SVR. In the case of alinearkernel, the parameters(epsilon)and C (cost)aredetermined using theGrid Search algorithm. In the Dated-Brent variable, the best parameter value is obtained with the value of C = 100 and = 1 while for the Alpha variable, the best parameter value for the SLC crude oil type is C= 0.01 and = 0.01, SC value C = 10 and = 1, Banyu value C = 100 and = 0.1, Banyu value C = 100 and = 0.1, Belida value C = 0.01 and = 0.1, Attaka value C = 0.1 and = 0.01 and Duri value C = 1 and = 1.TheAlpha value of the main crude oil type is the Duri crude oil type with the lowest RMSE value of 0.9651. The MAPE value for SC crude oil type = 19.55% and Duri = 19.46% is in the good category. The R2 value for Banyu crude oil = 0.60610, SC = 0.42717 and Duri = 0.50421 is in the good categoryand the MAPE value for Dated-Brent of49.73% is included in the faircategory

Des Suryani1, Mutia Fadhilla2

https://jurnal.iaii.or.id/index.php/RESTI/article/view/5551/900

nformatics Engineering, Islamic University of RiauPekanbaru, Indonesia

18-02-2024

FAJAR BAGUS W

PDF

ENGLISH

TEXT