Forecasting of Long-Term Natural Gas Consumption Using Linear Regression Method: A Study on Natural Gas Sector in Bangladesh

Abstract

Forecasting of natural gas consumption for a country is essential for its logistics and future work flow. Long-term oriented natural gas consumption is also important for efficient logistics management in the power and energy sector. It is an advanced process for a country or a city to know the future demand of gas that can improve the transportation system, especially when the gas is transported through the pipelines. In this paper, linear regression method is proposed to forecast long-term natural gas consumption in Bangladesh. This method is very useful in time series data and forecasting on long term basis. The trend line of the regression can forecast the long-term based demand of the natural gas in the country. The results of this study indicate that natural gas consumption in the country is increasing over time. Gas consumption has increased rapidly in the last 5-10 years and it is forecasted that the consumption will be increased drastically for the next few years. Some statistical parameters such as MAE, MAPE, and RMSE are performed to evaluate the forecasting accuracy. It is also seen that the regression method provides good results in both cases annual and sector wise analysis. The results of this paper show that the linear regression model is capable of forecasting long-term natural gas consumption in Bangladesh.