MODELING AND FORECASTING OF ELECTRICITY CONSUMPTION USING ARMA AND EXPONENTIAL SMOOTHING
The various available literature on modeling electricity consumption use time series models such as Autoregressive Moving Average (ARMA), Autoregressive Integrated Moving Average (ARIMA), Vector Auto-regression (VAR) and so on. In this research, Autoregressive Moving Average (ARMA)and Exponential Smoothing processes are used to fit the monthly electricity consumption (mega watt hour) from January, 2011 to Dec, 2015 (a case study of Gazaoua power plant). Augmented Dickey Fuller (ADF) test was performed to check for presence of unit root. The work, based on the Autocorrelation function (ACF) and Partial Autocorrelation function (PACF) examines the forecast ability of varieties of Autoregressive Moving Average (ARMA) and Exponential Smoothing. The result shows that seasonal Autoregressive Moving Average of first order and Holt-winters Multiplicative satisfied the best model criteria. However, it is further observed based on loss function that Holt-winters Multiplicative emerged as the best forecasting method for the data set.