This study develops a short term forecasting model for household electric power consumption in Seoul, which can be used for the effective planning and control of utility management.
The model developed is based on exponentially weighted moving average model and incorporates monthly average temperature as an exogeneous factor so as to enhance its forecasting accuracy.
The model is empirically compared with the Winters' three parameter model which is widely used in practice and the Box-Jenkins model know to be one of the most accurate short term forecasting techniques.
The result indicates that the developed hybrid exponential models is better in terms of accuracy measured by average forecast error, mean squared error, and autocorrelated error.