Forecasting residential natural gas consumption in Egypt

Abstract

Purpose – This paper aims to obtain accurate forecasts of the hourly residential natural gas consumption, in Egypt, taken into consideration the volatile multiple seasonal nature of the gas series. This matter helps in both minimizing the cost of energy and maintaining the reliability of the Egyptian power system as well. Design/methodology/approach – Double seasonal autoregressive integrated moving average-generalized autoregressive conditional heteroskedasticity model is used to obtain accurate forecasts of the hourly Egyptian gas consumption series. This model captures both daily and weekly seasonal patterns apparent in the series as well as the volatility of the series. Findings – Using the mean absolute percentage error to check the forecasting accuracy of the model, it is proved that the produced outcomes are accurate. Therefore, the proposed model could be recommended for forecasting the Egyptian natural gas consumption. Originality/value – The contribution of this research lies in the ingenuity of using time series models that accommodate both daily and weekly seasonal patterns, which have not been taken into consideration before, in addition to the series volatility to forecast hourly consumption of natural gas in Egypt.

Keywords

• egypt
• arima
• garch
• time series models
• multiple seasonality pattern
• post-sample forecasts