The Forecasting of Iran Natural Gas Consumption Based On Neural-Fuzzy System Until 2020

Abstract

In this paper, an Adaptive-Network-based Fuzzy Inference System (ANFIS) is used for forecasting of natural gas consumption. It is clear that natural gas consumption prediction for future, surly can help Statesmen to decide more certain. There are many variables which effect on gas consumption but two variables that named Gross Domestic Product (GDP) and population, are selected as two input variables. The input variables data and output variable (gas consumption) data are collected in years 1993 till 2012. Pre-process is done on the primary data to obtain better results then finally our outputs are post-processed. In this paper, many fuzzy models are applied and the results and error of every model are investigated. All ANFIS outputs are compared with real output by considering Mean Absolute Percentage Error (MAPE). The best model, which has the lowest MAPE, is chosen for forecasting gas consumption. In this paper, the values of gas consumption are forecasted from 2013 to 2020