Peramalan Harga Minyak Mentah Menggunakan Model Autoregressive Integrated Moving Average Neural Network (ARIMA-NN)

Abstract

Crude oil prices can affect the production and consumption of a country. Crude oil prices are set every month and semester by Indonesian Crude Oil Price(ICP). Bontang Return Condensate (BRC) is one of crude oil type in Indonesia. Forecasting is an expectation of a request or thing that will come based on several forecasting variables.In this study, data was used in July 2010-December 2017. The purpose of this study was to determine the best model of Indonesian crude oil price data for the BRC type and the forecasting results. The model used in this study is the ARIMA-NN model which is a combination of ARIMA model and Neural Network (NN) model. The best ARIMA-NN model has ARIMA (2,1,0) and NN components with 2 inputs and 2 neurons in the hidden layer. The NN model is a Feed Forward Neural Network (FFNN) model with backpropagation algorithm. The results of the forecasting of the BRC Indonesia crude oil price for January-December 2018 are around the value of 60 USD / Barrel.