Electricity Price Forecasting Using Neural Networks and Similar Days

Abstract

This chapter focuses on day‐ahead forecasts of electricity price in the PJM market using artificial neural network (ANN) model based on the similar days (SD) method. The PJM competitive market is a regional transmission organization (RTO) that plays a vital role in the US electric system. The chapter contributes to forecast electricity prices in the day‐ahead market. In addition to the integration of SD and ANN method, it also proposes a new technique to forecast hourly electricity prices in the PJM market using a recursive neural network (RNN), which is based on the SD method. The proposed RNN model is also applied to generate the next three‐day price forecasts. To evaluate the performance of the proposed neural networks, the mean absolute percentage error (MAPE), mean absolute error (MAE), and forecast mean square error (FMSE) are calculated.