Load forecasting - COVID period

Description

Data providad here is originated by REE (actual Spanish load), UMH (forecasts for Spain) and ENTSO-E platform (actual and forecast data from Germany. These files are associated with the research paper: Impact of COVID-19 Lockdown in Short-term Load Forecasting.

Steps to reproduce

datos_españa.mat contains 15-minutes (96 values) daily forecasts for 2019 and 2020 (731 days) . AR means autoregressive model, NN mean neural network model and C means combined model. Actual load is stored in matrix named real. Binary vector dia_especial is True if the corresponding day was considered special. analisis_alemania.m and Guion_articulo_masdatos.m contain the code to reproduce the results. The file Data_Dem_GE.mat contains the actual load from Germany. The file Data_Pred_GE.mat contains the reported forecast from Germany. The mean is not zero. For the purpose of the research the data is shifted to have zero-mean.

Institutions

Categories

Funding

Licence