A study of risk reduction for daily peak load demand forecasting

In the daily peak load demand forecasting, it is necessary to suppress the dispersion of the prediction result that becomes a risk as well as the improvement of prediction accuracy. Therefore, this paper aims to improve accuracy and risk reduction. This proposed method has calculated new predicted value using multiple regression model and Artificial Neural Network (ANN) in order to realize risk reduction. We applied the proposed method to actual electric power demand data of Tokyo electric power company. In verification of daily peak load demand forecasting, we confirmed that the proposed method has improved the dispersion of the prediction result.