Answers are grounded in this paper and cite their sources.
This paper examines specific features of electricity as a commodity and analyzes the subsequent difficulties arising when accurately forecasting the clearing wholesale electricity price. Statistical methods (different kinds of linear regression), machine learning algorithms (random forest, support vector machines, etc.) and deep learning models (perceptrons, recurrent neural networks) are applied for forecasting the supplier’s clearing electricity price for the 2nd price zone of the Russian wholesale electricity and capacity market. The forecasting results are evaluated using statistical error metrics such as R2, MSE, MAE, and others. The complexities arising from the volatile and nonlinear nature of electricity prices, as well as the challenges in comparing prediction models due to diverse datasets and evaluation metrics are addressed. The need to establish an open platform for exchanging forecasting methods and datasets, including electricity price and auxiliary features, is substantiated. The open platform will also facilitate cooperation among researchers worldwide to evaluate and enhance the accuracy and competitiveness of short-term electricity price forecasting methods.
No takes yet. Share an insight, caveat, or question.
Рожков et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: