Ensemble of relevance vector machines and boosted trees for electricity price forecasting

Abstract Real-time prediction of electricity pricing signals is essential for scheduling load demand in price-directed grids. In a deregulated electricity market, this helps substantially increase the gains of utility companies and minimize the electricity cost to the consumers. This paper introduces a novel model for electricity locational marginal price forecasting primarily centered on relevance vector machine. Two different versions of relevance vector machine are used based on Gaussian radial basis function and polynomial kernels in the first stage. The performance of the model is boosted using Extreme Gradient Boosting to incorporate the stochastic changes in prices. In the second stage, the outputs of the three models are stacked using Elastic net regression and the final price is forecasted after bagging the computed values. The model is trained and tested on real-time data of New England electricity market. Specifically, data for two years from 2012 to 2013 have been collected with a resolution of one hour. The proposed model has proven to be highly accurate and computationally cheap at the same time. It has been compared with various models that have been previously proposed for electricity forecasting including relevance vector machine, multilayer perceptron, random forest regressor, support vector machine, recurrent neural network, and least absolute shrinkage and selection operator. The proposed model is found to outperform all the other mentioned models with a mean absolute error of 2.6 on the test set and is sufficiently cheap computationally with a training time of 88 s.

[1]  Nikolaos G. Bourbakis,et al.  Genetic Optimal Regression of Relevance Vector Machines for Electricity Pricing Signal Forecasting in Smart Grids , 2015, IEEE Transactions on Smart Grid.

[2]  Jing Shi,et al.  Applying ARMA–GARCH approaches to forecasting short-term electricity prices , 2013 .

[3]  Reza Kheirollahi,et al.  Point and interval forecasting of real-time and day-ahead electricity prices by a novel hybrid approach , 2017 .

[4]  T. Dillon,et al.  Electricity price short-term forecasting using artificial neural networks , 1999 .

[5]  N. Amjady.  Day-ahead price forecasting of electricity markets by a new fuzzy neural network , 2006, IEEE Transactions on Power Systems.

[6]  Jakub Nowotarski,et al.  An empirical comparison of alternate schemes for combining electricity spot price forecasts , 2013 .

[7]  T. W. Berrie,et al.  Electricity Economics and Planning , 1992 .

[8]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[9]  Richard M. Everson,et al.  Smooth relevance vector machine: a smoothness prior extension of the RVM , 2007, Machine Learning.

[10]  J. Contreras,et al.  Forecasting next-day electricity prices by time series models , 2002 .

[11]  Bart De Schutter,et al.  Forecasting spot electricity prices Deep learning approaches and empirical comparison of traditional algorithms , 2018 .

[12]  N. Amjady,et al.  Day-Ahead Price Forecasting of Electricity Markets by Mutual Information Technique and Cascaded Neuro-Evolutionary Algorithm , 2009, IEEE Transactions on Power Systems.

[13]  Ioannis P. Panapakidis,et al.  Day-ahead electricity price forecasting via the application of artificial neural network based models , 2016 .

[14]  L. S. Moulin,et al.  Confidence intervals for neural network based short-term load forecasting , 2000 .

[15]  Michael E. Tipping.  Sparse Bayesian Learning and the Relevance Vector Machine , 2001, J. Mach. Learn. Res..

[16]  Hossam Mosbah,et al.  Hourly Electricity Price Forecasting for the Next Month Using Multilayer Neural Network , 2016, Canadian Journal of Electrical and Computer Engineering.

[17]  José Mira-McWilliams,et al.  Important variable assessment and electricity price forecasting based on regression tree models: classification and regression trees, Bagging and Random Forests , 2015 .

[18]  J. Contreras,et al.  ARIMA models to predict next-day electricity prices , 2002 .

[19]  Chen-Ching Liu,et al.  Day-Ahead Electricity Price Forecasting in a Grid Environment , 2007, IEEE Transactions on Power Systems.

[20]  B. Eswara Reddy,et al.  A moving-average filter based hybrid ARIMA-ANN model for forecasting time series data , 2014, Appl. Soft Comput..

[21]  M. Delimar,et al.  Dynamic Hybrid Model for Short-Term Electricity Price Forecasting , 2014 .

[22]  Alma Y. Alanis,et al.  Electricity Prices Forecasting using Artificial Neural Networks , 2018, IEEE Latin America Transactions.

[23]  Florian Ziel,et al.  Forecasting Electricity Spot Prices Using Lasso: On Capturing the Autoregressive Intraday Structure , 2015, IEEE Transactions on Power Systems.

[24]  T. Senjyu,et al.  Several-hours-ahead electricity price and load forecasting using neural networks , 2005, IEEE Power Engineering Society General Meeting.

[25]  R. Weron.  Electricity price forecasting: A review of the state-of-the-art with a look into the future , 2014 .

[26]  Zhang Yang,et al.  Electricity price forecasting by a hybrid model, combining wavelet transform, ARMA and kernel-based extreme learning machine methods , 2017 .

[27]  Narayanan Kumarappan,et al.  Day-Ahead Deregulated Electricity Market Price Forecasting Using Recurrent Neural Network , 2013, IEEE Systems Journal.