A1452
Title: oraklE$\_$R: An R package for long-term energy demand forecasting
Authors: Simone Maxand - Europa-Universität Viadrina (Germany) [presenting]
Johannes Schwenzer - Europa-Universität Viadrina Frankfurt(O) (Germany)
Tatiana Grandon - Uni Flensburg (Germany)
Abstract: An open-source R package is presented, which provides long-term (yearly), mid-term (daily), and short-term (hourly) energy load forecasts for all EU countries and beyond. The proposed load forecasting method is based on hybrid multiple regression models and long short-term memory (LSTM) for residual prediction. The package includes functions for the automatic loading of covariates, i.e., historic hourly load data and meteorological and macroeconomic data. It produces yearly load forecasts with hourly resolution for 35 EU countries with an average of 95\% accuracy. The flexible handling using the user's own database provided and predicted data allows for scenario building of future energy loads.
