Analysis and Forecasting of Demand for Electricity Using Time Series Analysis

Summary

Nearly all papers concerning short-term electricity forecasting are using either an “econometric” or an univariate approach. The study to be described shows that a combined model with exogeneous variables (especially weather variables) as well as dynamic information is most convenient for the analysed series. As for concrete examination, linearized regression models with autoregressive residuals turned out to be adequate. Applying autoregressive transformations we get equations which are easy to estimate, but nevertheless show good results. This article will point out some applications of time series analysis as well as self-reliant methods and approaches. As for theoretical background, refer to basis literature.