Billé, A.G., Gianfreda, A., Del Grosso, F., and Ravazzolo, F. (2023) embark on a comprehensive examination of electricity price forecasting methods in their research. Utilizing data from the Italian power exchange market, specifically, day-ahead electricity prices, they innovatively deconstruct the complex problem of electricity price prediction into several more manageable sub-problems. Each sub-problem is resolved via a hybrid approach, using a combination of expert models, linear models, and nonlinear models.
The authors divided the electricity forecasts into two steps. In the first step, they utilized statistical and machine learning techniques to create expert models for different subprocesses. In the second step, they applied linear and non-linear models for forecasting.
They tested multiple models including the ones based on time-series analysis methods (ARIMA, Holt-Winters), and more modern approaches, such as support vector machines (SVM), cubic spline interpolation, and artificial neural networks (ANN).
Through their study, the authors discovered that no single model consistently outperforms the others - the prediction accuracy varies depending on the specific circumstances.
However, the model combination approach always demonstrated promising performance across all situations. This practice of combining different forecasting methods is also known as "forecast combination".
Like all studies, this research has its limitations. Fundamental factors such as power plant outages, transmission constraints, fuel prices and weather forecasting errors were not considered in the models and could potentially impact the forecasting results.
Despite the limitations, the paper proposes a robust method for electricity price forecasting that provides more accurate forecasts than a single model approach.
Based on these points, And suggested data analysis and interpretations. The paper promotes a methodology that could significantly contribute to current practices in forecasting the electricity market prices in future research and practical applications.
Billé, A.G., Gianfreda, A., Del Grosso, F. and Ravazzolo, F., 2023. Forecasting electricity prices with expert, linear, and nonlinear models. International Journal of Forecasting, 39(2), pp.570-586.
Subscribe to our newsletter to stay up to date and receive our updated forecasts with an in-depth analysis every month.
.png){kind=logo}
