Poprawa jakości prognoz probabilistycznych cen energii elektrycznej na rynku dnia 
następnego przy użyciu metody głównych składowych
[Enhancing probabilistic forecasts of day-ahead electricity prices with principal component analysis (PCA)]

Tomasz Serafin (Doctoral School - Management, PWr, Wrocław)


Recent changes in electricity markets amplified the importance of accurate 
probabilistic forecasts. In this study, we apply principal component analysis (PCA) 
in the process of obtaining probabilistic forecasts of day-ahead electricity prices. 
We additionally focus on the impact of different data normalization schemes on the 
accuracy of forecasts. The performance is evaluated across two datasets from major 
power markets – Scandinavian Nord Pool and American PJM. The proposed approach is 
able to outperform well-performing literature benchmarks in terms of the empirical 
coverage of prediction intervals.