Krótkoterminowe probabilistyczne prognozowanie zapotrzebowania na energię elektryczną z wykorzystaniem GAMLSS
[Short-term probabilistic load forecasting with GAMLSS framework]

Katarzyna Chęć (Doctoral School - Management, PWr, Wrocław)


Forecasts of electricity demand have a wide range of applications in electricity price forecasting. However, 
the commonly available forecasts of the Transmission System Operators (TSO) can be of low accuracy and 
biased. It has already been shown in the literature that these forecasts can be improved with simple 
regression models, but this issue has only been the subject of research in the area of point forecasts. 
Therefore, the aim of this study is to investigate the application of probabilistic forecasting methods to 
obtain high quality electricity load (or demand) forecasts. The study focuses on the use of distributional 
regression models, more precisely - Generalized Additive Models for Location, Scale and Shape (GAMLSS).