New robust method applied to short-term load forecasting

In this paper, the stochastic characteristics of the electric consumption in France are analyzed. It is shown that the load time series exhibit lasting abrupt changes in the stochastic pattern, termed breaks, which need to be accounted for during the modeling process. Thus, a new robust diagnostic approach for which the identification of the breaks is carried out via a robust autocorrelation function estimates is introduced. The developed procedure consists of the following steps: (i) estimate the parameters of a high order autoregressive AR(p^*) by means of the ratio of medians, (ii) execute a robust filter cleaner to reject the outliers, and (iii) apply a maximum-likelihood estimator defined at the Gaussian distribution to handle missing values. The performance of this method has been evaluated on the French electric load time series in terms of execution time, ability to detect and suppress outliers, and forecasting accuracy. The new approach improves the load forecasting quality for rdquonormal daysrdquo and presents several interesting properties such as good robustness, fast execution, simplicity and easy on-line implementation. Finally, a simple vector time series method is proposed in order to deal with heteroscedasticity.