Short-Term Electricity Load Forecasting Model Based DSARIMA

Forecasting short-term electrical load is very important so that the quality of power supplied can be maintained properly. The study was conducted to measure results forecasting based on parameter estimates and presentation time series data. It manage stationary data, both in terms mean variance. Data done by determining value variance through Box-Cox transformation method ACF PACF plots. This considers pattern electricity consumption which contains double seasonal patterns. previous studies show electric prediction model, DSARIMA model with a MAPE 2.06%. condition used predict still has tendency not normally distributed it estimated there are outliers. Improvements AR MA parameters meet standard error tolerance 5 percent increased this study. showed improvement model. significance obtained 1.56 when compared actual