Daily electricity demand forecasting in South Africa

摘要

The paper investigates the impact of day of the week, holidays and other seasonal effects on daily electricity demand in South Africa using regression, seasonal autoregressive integrated moving average (SARIMA) and regression model with SARIMA (RegSARIMA) models for the period 2001 to 2009. The results from this study show that the SARIMA model produces better forecast accuracy with a mean absolute percent error (MAPE) of 1.36%. The RegSARIMA model had a MAPE of 1.75%. From these model results, we can conclude that, holidays play a major role in determining the demand of electricity.