IMPROVEMENT OF SHORT-TERM ELECTRIC LOAD FORECAST

Abstract

In this paper, the author has proposed a two-step power demand forecasting algorithm. In the first step, the algorithm with the best accuracy among the selected supervised learning prediction algorithms is selected. In the second step, the authors propose a method to calculate an offset to add to the forecast value with the desire to reduce the negative deviation and improve the accuracy of the method. The testing has showed the effectiveness of the improvement proposals