Evaluasi Prediksi Konsumsi Gas Bumi Menggunakan Artificial Neural Network (Ann)

Abstract

The national energy demand, especially the oil and gas sector, is increasing in line with the increasing population and the condition of national economic growth which continues move positively. The increase of energy demand is on average more than 5% per year for this decade. Meanwhile, the condition of national oil and gas reserves and production sector continues to decline every year. This has resulted in Indonesia becoming a net importer of oil and gas. Domestic demand for natural gas increases every year, while on the other hand Indonesia still has commitments to sell natural gas abroad, pipeline gas and LNG. For this reason, a more accurate prediction of natural gas in Indonesia will be very helpful for policy makers so that policies taken are right on target so that natural gas which should be consumed domestically is not exported abroad. One of the good prediction methods is using artificial neural network (ANN). In this study, the input data used are economic growth, population, and gas prices, while the output data is natural gas consumption. This study uses five ANN architectural models that are formed. From the simulation results, the best accuracy is model 1 with an accuracy of 96.89%.