Electricity price forecasting using Support Vector Machines by considering oil and natural gas price impacts

Accurate electricity price prediction is one of the most important parts of decision making for electricity market participants to make reasonable competing strategies. Support Vector Machine (SVM) is a novel algorithm based on a predictive modeling method and a powerful classification method in machine learning and data mining. Most of SVM-based and non-SVM-based models ignore other important factors in the electricity price dynamics and electricity price models are built regard to just historical electricity prices; However, electricity price has a strong correlation with other variables like oil and natural gas price. In this paper, single SVM model is used to combine diverse influential variables as 1-Historical Electricity Price of Germany 2-GASPOOL price as first natural gas reference price 3-Net-Connect-Germany (NCG) price as second natural gas reference price 4- West Texas Intermediate (WTI) daily price as US oil benchmark. The simulation results show that using oil and natural gas prices can improve SVM model prediction ability compared to the SVM models built on mere historical electricity price.

[1]  Liang Tian,et al.  A novel approach for short-term load forecasting using support vector machines , 2004, Int. J. Neural Syst..

[2]  Nello Cristianini,et al.  An Introduction to Support Vector Machines and Other Kernel-based Learning Methods , 2000 .

[3]  Wei Sun,et al.  Application of Time Series Based SVM Model on Next-Day Electricity Price Forecasting Under Deregulated Power Market , 2006, 2006 International Conference on Machine Learning and Cybernetics.

[4]  Marco van Akkeren,et al.  A GARCH forecasting model to predict day-ahead electricity prices , 2005, IEEE Transactions on Power Systems.

[5]  M. Hildmann,et al.  A quantitative analysis of weather effects on traded volume in the Swiss energy spot market , 2012, 2012 9th International Conference on the European Energy Market.

[6]  Li Jinying,et al.  Next-day electricity price forecasting based on support vector machines and data mining technology , 2008, 2008 27th Chinese Control Conference.

[7]  J. Contreras,et al.  Forecasting Next-Day Electricity Prices by Time Series Models , 2002, IEEE Power Engineering Review.

[8]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[9]  A. Gil,et al.  Forecasting of electricity prices with neural networks , 2006 .

[10]  R. A. Swief,et al.  Support vector machines (SVM) based short term electricity load-price forecasting , 2009, 2009 IEEE Bucharest PowerTech.

[11]  Guolian Hou,et al.  Forecasting next-day electricity prices with Hidden Markov Models , 2010, 2010 5th IEEE Conference on Industrial Electronics and Applications.

[12]  Zuyi Li,et al.  Adaptive short-term electricity price forecasting using artificial neural networks in the restructured power markets , 2004 .

[13]  Stein-Erik Fleten,et al.  Modeling Long-Term Electricity Forward Prices , 2009, IEEE Transactions on Power Systems.

[14]  Wei Sun,et al.  Application of neural network model combining information entropy and ant colony clustering theory for short-term load forecasting , 2005, 2005 International Conference on Machine Learning and Cybernetics.

[15]  J. Contreras,et al.  ARIMA Models to Predict Next-Day Electricity Prices , 2002, IEEE Power Engineering Review.