This repo holds the experiments for the Paper Online Multivariate Regularized Distributional Regression for High-dimensional Probabilistic Electricity Price Forecasting (Link to Arxiv).
The following packages are needed to run the experiments
- scipy
- numba
- numpy
- pandas
- scikit-learn
- rolch
- scoringrules
- matplotlib
- seaborn
- autograd (for testing)
The data in experiments/epf_germany is taken from Marcjasz, Grzegorz, et al. "Distributional neural networks for electricity price forecasting." Energy Economics 125 (2023) and the according Github repository.
Please ensure you can import the code from /src. This is done in the experiment code by temporarily appending "..\..\..\online_mv_distreg" to the PATH for some of the Python files. If you're on a Windows machine, you might need to adjust this code. The code therein holds the estimator classes for the multivariate online distributional regression model.
The forecasting study is in experiments/epf_germany. Please run the file 00_run_study.py to run all experiments in one go.
If you run the files separately, please keep the order for files 01 to 05 inclusive.
Simon Hirsch is employed as an industrial PhD student by Statkraft Trading GmbH and gratefully acknowledges the support and funding received. This work contains the author’s opinions and does not necessarily reflect Statkraft’s position.